Msc. Thesis-Computer Science 1. USMAN, GAMBO ABDULLAHI PGS/00295 MODELLING AND SIMULATION STUDIES OF ARTIFICIAL PANCREAS 1994 2. HAROUNA, NAROUA PGS/00321 DEVELOPMENT OF A SOFTWARE TOOL FOR THE EVALUATION OF STRUCTURALISM IN dbase PROGRAMS 1994 The application of the methods of mathematical modelling, dynamic systems analysis and simulation is playing an increasingly important role in the study of biological systems in general, and metabolic and endocrine processes, both in physiology and in clinical medicine. The proper use of such techniques usually provided the physician, medical or biological research worker in the life sciences, with a greater understanding of the nature and behaviour of the complex processes that occur in physiology. Equally, mathematical models are now increasingly being used as aids in the diagnosis and treatment of diseases. This study essentially looked at Diabetes Melitus by studying the models of the Artificial Pancreas. Real world data about diabetic patients were collected from some hospitals and processed. A simulation experiment using the GPSS Simulation Language was designed and implemented. The results obtained from the Computer Simulation of two alternative control algorithms for the artificial pancreas, namely, the Albisser and the Biostator-Miles algorithms, indicated that the Biostator algorithm performs better than the Albisser algorithm in normalising the diabetic patient. It was also discovered that the route of insulin administration plays an important role in the normalisation of the metabolic state of a diabetic patient. This is indicated by the performance of the portal route as compared to the peripheral route. The art of programming seems to be full of elusive goals. There is no doubt that software can be produced in an efficient, timely fashion. It is known to all of us and we have even tried it. But, we have always asked ourselves the following questions: Is our program going to fail? When and how is our program going to work? A serious work has been done after that goal. Brilliant people have 3. MUKTAR, LIMAN AHMED PGS/00296 STUDIES ON FILE ORGANISATION TECHNIQUES: A CRITICAL AND COMPARATIVE ANAYSIS 1994 toiled for years to find the answers to those questions and they have produced masterpieces. For that same purpose, a complete static analysis was carried out on the Abubakar Tafawa Balewa University Computer Science final year students projects. The software engineering environment was restricted to programs written in dBase. To prove a program correct, it was tested for structuralism using the static analysis of program segments. Both theory and algorithms were developed on structuralism. Finally, a package was developed to assist the analysis by providing statistics on the program structure based on program reduction. It was observed that the students made good use of SESE (SINGLE ENTRY, SINGLE EXIT) constructs. It was also observed that the students made use of dBase control and data structures. The major source of students problems was found to be the use of LOOP statement. Five file organisation techniques were studied, analysed and compared. These techniques include Sequential, Indexed-sequential, Direct, Multi-indexed, and Multiring indexed, and Multring files. Each of these methods is studied and analysed based on the data transfer rates, average seek times of some direct access devices. Some hardware parameters were derived and in turn performance formulas were written for each technique. Simulation run was carried out for each method. The performance was measured in terms of record size, time to fetch a record, get-next record, insert a record, update a record, read entire file and file reorganization. Simulation results show that when a fixed record size is used in an organisation, sequential and multi ring files require the minimum bytes to store a record hence the best of all the methods. While Indexed-sequential is the fastest in fetching a record as well as in getting the next record, sequential file is the fastest in insertion and updating of record as well as in file reorganization. The 4. LAWAN, ABDULWAH AB 5. AKINYEMI, ADESINA ALABA PGS/00688 APPLICATION OF COMPUTER SIMULATION IN MULTIPLE SERVER FINITE QUEUEING SYSTEM. 1996 ESTIMATING SOFTWARE COST IN A DEVELOPING 1997 direct method placed the second fastest method in fetching and updating of a record as well as in file reorganization. Indexed file organisation performs fairly good in fetch, get-next and insertion of a record. From the analysis it was found that sequential file organization gave overall best performance then followed by indexedsequential. Direct method came third while the fourth position went to indexed file organization. Multiring method took the last position. This project presents a computer simulation approach, which evaluates the performance criteria of a multiple server finite queueing system for different number of servers, with a view to determining through comparison of the result which number of servers gives the optimum of the performance criteria. Texaco filling station, Jos road, Bauchi was used as a case study. The arrival rate follows a poisson process, and the service time are exponentially distributed. The performance criteria are measured in terms of average number of vehicles in queue and in system per minute, average waiting time in queue and in system per vehicle, average utilisation of service facility, average number of vehicles serviced per day and average contribution made as a profit per service point. The starting data required for the simulation experiment is obtained from the filling station. The data collected provided necessary inputs for the simulation model. The simulation result showed that the number of service points that produced the optimum of the specified performance criteria of the system was three. These, inspite of the fact that the chances of three service points being idle was increased, this increase was insignificant when compared with the benefits derived from the system. Estimating the cost of a software product is one of the most difficult and error-prone tasks in software engineering. The dynamic nature of programming and COUNTRY: NIGERIA AS A CASE STUDY 6. YUSUF, BASHIR KAGARA PGS/00974 SIMULATION OF KERNEL MULTITASKING OPERATING SYSTEM IN UNIX ENVRONMENT. 1997 the need for objective pricing system for software products are pre-requisite for the assessment of different methods of software estimation. This study looked at different methods of estimating cost of software with a view to arriving at an acceptable and widely used method. An experimental program using Turbo PASCAL language was designed and implemented for constructive cost model (COCOMO) and Putnam cost estimation model. The performance was measured in terms of man hour, programmer experience, duration, staff size, number of source codes involved, reliability and lines of documentation. The results obtained from the Program of software cost estimation usmg constructive cost model (COCOMO) vs PUTNAM cost estimation model methods show that it is cheaper using COCOMO than using Putnam cost estimation model method. It was observed that software developers may prefer COCOMO method to Putnam cost estimation method and other methods, but may unconsciously make use of expert judgement and delphi cost estimation methods interchangeably. It was also observed that in a developing country the efforts of Analysts and programmers are not always well rewarded and recognized, since, their products are assumed to be inferior to the imported packages which is not always true. A general way to make a smaller model of a large system, or to represent the fact that the observations possible on it are limited, is to apply an abstraction to it. simulation techniques have a far reaching impact on digital computer systems in general, and kernel multitasking operating system implementation is one area that deserves critical study for efficient management of computer system resources. The study was an attempt to model the actual kernel of UNIX operating system and evaluated various design options and trade-offs as they 7. BOUKARI, SOULEY PGS/00970 PERFORMANCE EVALUATION OF DIGITAL COMPUTER SYSTEM 1998 were encountered. A simulation model, using C++ programming language, was designed and implemented. The maj or functions of the Kernel studied include process management, interprocess communication and synchronization, interrupt management and startup of the initial system configuration. The model was implemented on GATEWAY 2000 486 microcomputer running on sco UNIX Release 3.2 Version 4.2 operating system. Numerical experiments on random instances of the performance metrics suggest that the model performs satisfactorily on the average. A classical concurrency problem, the dining philosophers problem, was solved using the model. With the rapid involvement of industries and many business oriented organisations in computing, a tool for evaluating, analyzing and planning computing systems becomes imperative. For most computer systems, response time, resource utilization and throughput are the primary parameters of performance. In order to evaluate and predict the above parameters and other aspects of performance of a digital computer system, it becomes necessary for us to represent the compute systems and its workload by some models. This works presents some tools for evaluating the performance of digital computer system with a view to determining through comparison of the results obtained which approach gives the best optimum performance. NITEL, North-East Zone and A. T. B. U Computer Centen, Bauchi, Nigeri a were considered as case studies. The starting data required for the simulation were obtained from NITEL MIS Data Processing Unit and A.T. B. U. Computing Unit. The data collected were used as input for the simulation model. The results show that the mean response time and the mean throughput of the system obtained from the basic simulation model are more realistic than the ones obtained with queuing model. 8. ALI, JOAN BERI PGS/00833 SOFTWARE RELIABILITY MEASURES USING SOME RELIABILITY MODELS 1998 9. TITIKUS, ISHAYA AUDU PGS/96/01 131 A SOFTWARE FOR COPING WITH UNCERTAINTIES WITH APPLICATION TO YOLA-ADAMAWA STATE AGROCLIMATOLOGY 2000 Simulation model is therefore recommended as a powerful tool for evaluating and predicting the performance measures of digital computer systems. This work uses reliability models to estimate the reliability of ATBU payroll software. The models used are the Basic-Reliability, Jelinski- Moranda, SchickWolverton and the Path decomposition models. The results show that the Jelinski-Moranda and Path decomposition models with a reliability, R(t), of 8.6 x 10-1 and 7.03 x 10-1, respectively, performed better than the others. This corresponds to a Mean Time To Failure, MTTF, of 13.7minutes and 12.99minutes respectively. It was concluded that the Jelinski-Moranda and Pathdecomposition models could be used in the calculations of reliability for softwares that have similar characteristics with the ATBU payroll. One of the essence of knowledge is enabling human beings to cope with the odds of their environments, particularly if such odds can not be gotten rid off. One of such environmental odds is weather on which man so far has no control. The study presents how computer capability could be harnessed to assist man in taking decisions that will enable him cope with the odds of his environment. In this regard, Yola-Adamawa State as an example of areas where farmers battle with Agroclimatological uncertainties was studied. Data on rainfall, price list of crops, farm inputs and crops yield per hectare under various climatic conditions. were collected and analyzed to usher in a crops decision- matrix. The following uncertainty decision criteria were studied, presented and incorporated into the software: Maximax, Maximin, Minimax, Laplace, Hurwicz, Expected Monetary Value (EMV), Expected Opportunity Loss (EOL) and Expected Utility Value (EUV). Though, the most widely used uncertainty decision criterion is the EMV, yet it has a major shortcoming since it fails to 10. SULEIMAN, ABDULMALIK PGS/01249 A STUDY OF SOME ROUTING ALGORITHMS 2000 11. KWAMI, ABUBAKAR MOHAMMED PGS/0348 APPLICATIONS OF FOURIER TRANSFORM TO THE GENERALISED D’ALERMBERT WAVE EQUATION. 2000 address the decision-maker's personal attitude to risk. This inadequacy hitherto made this decision criterion less desirable. An Index of Variation (IV), which measures the risk potentials of each alternative, has been introduced in this project as a means of solving this inadequacy. This software, which is of general application to uncertainty decision- matrices, was then applied to the crops decision- matrix of the study area, and the system displayed the economic expectation and the risk status of each crop. Based on the economic expectation of the crops which is measured in terms of Expected Monetary Value (EMV), and their climatic risk status which, is measured in term of Index of variation (IV), the system identified respectively, Groundnuts and Guinea Com as the most viable crops to be grown in Yola-Adamawa State. In this project report, a survey was conducted on some routing algorithms. Based on this survey, an algorithm for channel routing using a two layer channel which finds an optimal solution in VLSI design was proposed. Associated with the channel routing problem are two constraints: Horizontal and Vertical. Using the horizontal and vertical constraints graphs, a searching tree was formed. Using a good dominance rules to terminate unneccessary nodes in the searching tree and a proper cost function, the algorithm was terminated and an optimal solution obtained as soon as a leaf node is reached. Experimental results show that the algorithm behaves quite well in average cases which coincides with theoritical analysis. Partial differential equations occur in some areas of science and engineering, and their applications to real life situation is of great .importance, therefore it is necessary that some method be employed in solving them. Although various methods such as separation of variables, method of characteristics, and laplace 12. DAPIAP, STEPHEN BOERWHOEN PGS/97/01 268 STUDY OF ALGORITHMS FOR DEADLOCK DETECTION 2000 transtorm method have been proposed and proved to be successful for certain class of equations, but their complexity and ambigity make it necessary for an effort co be devoted to some other methods of solution. As a result, much attention have been given to the use of Fourier transform method which is often simpler and widely accepted. The subject of this project report is Fourier transform method in the solution based of the study of on the generalised D'Alemgert- wave. equation.This equation was solved by Ducati and Capelas De Oliviera(In1998) using Riemann's method. Also, the solutions of special cases are Obtained using Fourier transform method. The result obtained using Fourier transform method is similar to that obtained using Riemann's method, and is shorter and simpler compare to the Riemann's method. Deadlock Detection, one of the three techniques for handling deadlock problem was selected and the algorithms to implement it were analysed and compared. The comparisons were made in order to measure their performances and determine the most efficient algorithm in terms of time and search paths for solving deadlock problem by deadlock detection. Breath-first search, Depth-first search, and the Horizontal and Vertical search were the algorithms analysed and compared. The algorithms were compared based on representing two resource allocation systems on directed graphs. The first system was simple with six process nodes and five resource nodes whereas the second system was more complex .with twenty process nodes and seventeen resource nodes. In each of the systems, three different situations based on resources requested by processes were considered: request that resulted in a deadlock situation, request for a free resource and its immediate allocation, and request for a resource not free yet waiting for it was deadlock-free. The programming language 13. AGAJI, IORSHASE PGS/01152 STUDIES IN 2000 COMPUTERBMEMO RY MANAGEMENT IN A NETWORK ENVIRONMENT 14. EKANEM, BASSEY ASUQUO PGS/9899/404028 COPING WITH SOFTWARE MAINTENANCE COST WITH RESPECT TO WARRANTY TERMS AND MANPOWER DEVEOPLMENT: A CASE STUDY OF CREATION NIGERIA LIMITED. 2002 C++ was used to write the programs for carrying out the required operations on a computer system. Amongst the three algorithms, the Horizontal and Vertical search algorithm performed better. The Depth-first search algorithm and the Breadth-first search algorithm in that order follow this respectively. The Horizontal and Vertical search algorithm was therefore found to be most efficient for implementing deadlock detection strategy. Memory management in a network environment was studied. A simulation program which minimises the cost and suggests the shortest path of page migration across the network by analysis a corresponding communication matrix was developed. This program was tested using star network of a commercial bank with twelve nodes across Nigeria. The results of the simulation program were compared to the minimum spanning tree of the network It was found that the cost of communication is minimal using the simulation program. This research work used Software Maintenance Cost Models and related Models to devise a means of coping with software maintenance cost for any given warranty term through careful planning of the release time and proper manpower deployment. Models used include, Warranty and Integration-Test, Wolverton PenaltyIncentive, Cost-Overrun Penalties, Mills model, MTTF model and Reliability models. Results of the research work conducted on Software Projects of Creation Nigeria Limi~ed showed that the best way to cope with the Project Management System (PMS) maintenance within the warranty period of 12 months is to release it at the end of the 7th week of its integration test. This will limit field errors to 24 as against other higher values, if released earlier. PMS reliability at the release time is 0.7. For Loans and Project Management System (LPMS) project, the optimum release time is at the end of the 8th week of integration. This limits field errors to 33 against 15. AHMADU, ASABE SANDRA PGS/98/40 4011 A COMPUTER BASED HUMAN RESOURCE INFORMATION SYSTEM A CASE STUDY OF FEDERAL UNIVERSITY TECHNOLOGY, YOLA. 2002 other higher values if released earlier. Extending the process beyond their respective optimum release time will definitely reduce the number of field errors but such extension is not tenable considering the high cost of additional integration test with the low cost of removing the field errors. Results also showed that the total development force will drop to 0 when all the 9 projects considered would have been released. This is due to the total workforce being gobbled up in maintenance of the released projects, a situation that requires urgent attention if the software organization must survive. Traditional human resources management systems rely on the intuition and whimsies of Human Resource Managers to take critical decisions that concern the selection, recruitment, development and utilization of human capital. In small, non-critical rudimentary industries, this system serves its purpose very well. However, in organizations where critical talents are needed, there is the need for highly responsive Human Resource Information System to aid today's Human Resource Managers to identify, select, develop and retain the talents needed to guide the operations of the organisation in this era of Information Technology. In this project a computer based Human Resource Information System is developed. The software provides a guide on how to effectively and efficiently manage human resources in a university system such as Federal University of Technology Yola. The software package provides inquiries on staff list by identity number, name, department, rank, state, Local Government Area, year spent and also the list and the count of staff in a particular job group (Junior or Senior), and job class (Academic or non Academic). It provides vital reports that help management in deciding who should be appraised, identify staff that requires training, identify skills that need to be developed, deploy staff from areas 16. OBEWA, OLUCHI JENNIE PGS/01188 COMPARISON OF ALTERNATIVE QUEUING SYSTEMS IN A BANK USING SIMULATION TECHNIQUES. A CASE STUDY OF FIRST BANK NIGERIA PLC. BAUCHI BRANCH 2002 of excess to areas of need thus enhancing good staff utilization. Visual Basic 6.0 provided a veritable platform for developing the front end using object oriented programming paradigm while microsoft access was used to creat and mount the Back end for enterprise connection as a database server. When limited service facilities fail to satisfy the demands for service that are made upon them, bottlenecks occur which generates queues or waiting lines. In business, queues have some "economic" or "cost" implications which include; cost of allowing long queues to build up, cost of speeding up services to reduce waiting times and queue lengths. Queuing problems are therefore concerned with minimizing average waiting times, the average length of queues, finding number of service points that ought to be used, the cost of servicing system and the benefits from cutting down waiting times. In this work, a software Bank_Queue_Simulator has been developed using Turbo Pascal for windows to simulate two queuing models, namely; the single-queue - multi teller model and the multi-queue - multi teller model. The results of this experiment show that using the single queue system running on five (5) tellers greatly reduces the waiting in the system and the length of the queue to 5.47 minutes and a customers in queue per minute, respectively. The cost of employing 5 tellers is less than that for the current 3- teller system. The probability of an arriving customer finding all tellers busy and that of finding an idle teller is fifty - fifty. For any meaningful reduction in the waiting times and queue lengths respectively to be achieved, the management is advised to consider employing 2 more tellers in addition to the existing 3- tellers to speed up service. Though this is subject to what management will consider reasonable wait time. It is further recommended that management should consider a re - training program for its teller staff on effective and efficient services. 17. OYONG, SAMUEL BISONG PGS/9899/404021 TIME ESTIMATION TOOL FOR SOFTWARE DEVELOPMENT PROJECTS. CASE STUDY: UNIVERSITY OF CALABAR COMPUTER CENTRE (Software Development Unit) 2003 18. EKABUA, OBETEN OBI PGS/9899/404031 SOFTWARE COMPOSITION AND BUILDING COMPLEX SYSTEMS: CASE STUDY OF DECK OIL SERVICES COMPANY LIMITED, PORTHARCOURT. 2003 This research work is aimed at improving the time for estimating new software projects. A tool was developed that would be used, given the set of general user requirements and technical requirements, to estimate the time to use and complete a given project. The tool made use of source lines of code which were determined in terms of size using statistical models, while Effort and Schedule were determined using Boehm et al (1995) COCOMO II Models. Putnam and Myers (1992) Software Life Cycle Management (SLIM) model was also used to determined the core metrics. The software development time was computed from the models developed by the two schools of thought: COCOMO II and SLIM. In practice, Size estimation and hence Effort and Schedule estimation are done either from experience (the Delphi method) or by Analogy. The researcher's contribution to knowledge is the development of a Tool that uses the theoretical method, (Algorithmic models), of determining the core metrics for new projects using software models. This tool was used or applied in The University of Calabar Computer Centre to estimate the time to use and develop Revenue collection program for Equity Bank Plc. It was found that it is good to use models from both schools of thought in developing a software project. No single model can be used to achieve a satisfactory result. One persistent problem in software engineering is how to put software systerns together out smaller sub-systems. The software systems of today are rapidly growing in number, sophistication, size, complexity, amount of distribution and number of usres. Software architectures promote development focused on modular building blocks (component), their interconnections (configurations) and their interactions (connectors). The emergence of software architecture and architectural style has focused attention on a new set of abstractions with 19. EDIM, AZOM EMMANUEL PGS/9899/404030 MODEL FOR SOFTWARE COST ESTIMATION: A CASE STUDY OF MULTISOFT NIGERIA LIMITED, PORT HARCOURT. 2003 20. ETIM, EDEM EKONG PGS/9899/404022 COMPUTER IMPLEMENTATION 2004 wltich we can create and compose software systems and gain intellectual control of software development. This thesis objective is to provide a model for the composition of different architectural styles within software systems and a disciplined approach to the process of architectural composition. The properties for reservoir formation (Oil, Gas, water, Rock) are used as different component (or modules) which are interconnecled togeUrer to enable evaluation of the reservoir formation. The advantage is that all the different components now "talk" to each other and share information. It is therefore recommended thai this software be used to enhance calculations involving reservoir formation evaluation. Software cost estimation is currently one of the most researched area in software engineering. The reasons being that software cost estimation provides the vital link between the general concepts and techniques of economic analysis and the particular world of software engineering. Several models and techniques have evolved within the last two decades. This research work reviews some of the existing models. The COCOMO II - multiple regression and Bayesian analysis models form the central focus of this study and have been presented in detail. "SCET" (Software Cost Estimation Tool) was developed to simulate the COCOMO II model as the main program, while Putnam (SLIM) and intermediate COCOMO were added as sub modules. Test data were collected from MULTISOFT SYSTEMS NIGERIA LIMITED. PORT HARCOURT. The data have been used to simulate results of effort, schedule and cost of projects in the mentioned organization. The results have been presented and discussed. The result predictions from the COCOMO II model seems more accurate than other models, this makes the model more preferable to others. Iterative methods for solving linear system of equations are interesting and valuable methods for solving linear OF THREE STATIONARY ITERATIVE METHODS FOR SOLVING LINEAR SYSTEMS. 21. ESIEFARIENR HE, MICHAEL BUKOHWO PGS/9899/404034 SOFTWARE PROTOTYPING SYSTEM ANALYSIS AND DESIGN (A CASE STUDY OF NIGERIA POLICE CRIMINAL DETENTION SYSTEM BAUCHI METROPOLIS) 2004 22. EKPENYONG, MOSES EFFIONG PGS/9899/404037 SIMULATING THE BEHAVIOUR OF PARALLEL MACHINES ON 2004 system of equations. In this project, three iterative methods were studied. Ten systems of linear equations were chosen for the implementation. The parameters used in the study were; accuracy of results, number of iterations, running time and storage. The results show that successive over relaxation performed better than Jacobi, and GaussSeidel methods. This relationship between System Analyst and Client (users) in software development is such that there are significant differences in their level of understanding of the problem and communication difficulty in expressing software requirement of the problem. These differences invariably lead to misconstrued system requirement, completely wrong software specification, late software deliveries, loss of man-hour and millions of Naira. This work advocates a client/Analyst involvement in system requirement elicitation and analysis using software prototyping techniques. This technique leads to the design of efficient software that meets the users requirement and save cost. In demonstrating this concept, the researcher developed a software prototype for Criminal Detection System using the Nigeria Police Crime Department, Bauchi State as a case study. These throw-away prototypes enable various concepts and design inherent in software prototyping to be fully discussed and demonstrated. Using the Prototype, it was observed that Criminals can easily be detected if the system is implemented by the Nigeria Police Force. It was also observed that the use of Prototyping in System analysis and Design saves time and money and thus lead to more efficient software where the client and developer are all involved in the development process. The efficiency of parallel processing is dependent upon the development of programming languages that optimize the division of tasks among the processors. This research simulates the behaviour of parallel machines on SEQUENTIAL ENVIRONMENT. 23. HASSAN, ANAH BIJIK PGS/20002002/4040 64 LOAD DISTRIBUTION IN CLIENT-SERVER IMPLEMENTATION CASE STUDY: MATERIALS MANAGEMENT IN STEYR NIGERIA. 2004 24. BLAMAH, NACHAMAD PGS/20002002/4040 THE RELIABILITY AND SECURITY 2004 sequential environment using java threads. Two parallel machine models were simulated: The Single Program Multiple, Data (SPMD) on arbitrary threads and Multiple Program Multiple Data (MPMD) on three threads, which are special cases of Single Instruction Multiple Data (SIMD) and Multiple Instruction Multiple Data (MIMD). The performance of these models were analysed and compared with a sequential model. It was observed that the runtime perfor-mance of the simulated SPMD indicated parallelism with increase in number of elements, and that of the MPMD was faster than the sequential version by 80ms. The MPMD results can further be improved by running programs with balanced load. In this work, client-server model is studied, and described as a system in which clients make requests while servers provide service. Design models have been considered, thin client has been praised in theory as the best. However, in practice, excess server load is identified as a major disadvantage in thin client-server implementation. This is because, the processing logic .and data are deployed on centralized servers. Fat-single executable (thick) client- server model is proposed to run on distributed database system. Steyr Nigeria Materials Management is considered as case study. Unified modeling language (UML) approach is used in designing the system. The proposed approach is implemented using Java programming Language because of its objectoriented nature and also it provides remote procedure call (RPC) in the form of remote method invocation (RMI). Simulation results are obtained for some operations for the thin and the thick client. They show that the proposed approach performs better and thus gives a better response time. This research considered online voting system as a critical system, where systems dependability is of 25. A VACHAKU 54 ONLINE VOTING SYSTEMS: A CASE STUDY OF THE INDEPENDENT NATIONAL ELECTORAL COMMISSION (INEC) OLUSANYA, O. MICHEAL PGS/9899/404025 COMPUTER APPLICATION OF MARKOV CHAIN MODEL ON TURBERCULOSIS OF THE LUNG. 2004 paramount importance. The most important dimensions of dependability that were considered are reliability and security. Two complementary approaches to systems reliability were used: fault avoidance and fault tolerance. Fault avoidance was ensured by the use of approaches to systems development that minimize faults before systems delivery. These include avoidance of dynamic memory allocation, multiple generalization, pointers, interrupts, recursion, and parallelism. Fault tolerance was achieved by implementing mechanisms for defensive (preventive) and retrospective fault detection, damage assessment, and forwardlbackward fault recovery. Two main architectures for fault tolerance implementation were considered. These are the N- version programming architecture (majority voting technique) and the recovery block architecture (redundant spare of codes.) Systems security was implemented by making each cast vote a function of the voter identification number (VIN) and by encrypting both the VIN and the vote, thereby making it impossible to detect who the voter is and who is being voted for. The unified modeling language (UML) was used for the design of the system, while lava's robust nature made it suitable for the implementation. Kiosk and remote online voting systems with distributed servers were used to implement both the reliability and security strategies. The resulting software was tested using sample data drawn from the case study and the results obtained were satisfactory. A suitable software architecture for the system with loosely coupled subsystems was finally presented. The computer application of Markov Chain model has been developed to project the probability of incidence of Tuberculosis (TB) in Bauchi State. Projections were made under the assumption that current TB control measures will remain unchanged for the periods. The program was tested using TB cohort report from Bayara 26. GOTENG, LONGINUS GOKOP PGS/20002001/4040 52 INTERGRITY CONSTRAINTS IN DISTRIBUTED DATABASE SYSTEMS: A CASE STUDY OF THE BOARD OF INTERNAL REVENUE. 2004 General Hospital, Bauchi, Bauchi State. The result shows the trend of TB incidence in 2001-2021 (of the population used) and it was discovered that at the end of the prediction, more than 80% of the population used might be affected. More so, the equilibrium state was checked with the same result and it was discovered that in next 65 years the probability of the Normal & Disease will tend to zero independently and Death will tends to one (L) if nothing is done to control it. The program was written in C++ and it can be used in any community to predict the probability incidence of TB. It accepts available data and prediction can be made. Distributed systems offer more flexibility, scalability, availability, reliability and speed than the old tightly secured centralized systems. Serious problems inherent with distributed systems such as security, update, replication and consistency of databases at different locations (sites) must be checked. This project explores constraints that enforce the integrity (consistency) of databases in distributed systems using semantic integrity assertions. Four steps were used to accomplish this: 1) accept input u, as an update; 2) send input to integrity constraints checker, c; 3) if valid, continue the process and send an output; 4) if invalid, keep it in the buffer for constraint checker reference and proceed to the next input. Built-in audit trail in operating system was used to track malicious updates from any site. The implementation of the integrity constraints using The Board of Internal Revenue as a case study was accomplished using Java programming language. This is because Java has features that can handle distributed programming more than most programming languages. An object- oriented approach was adopted in the design using the Universal Modeling Language (UML). Results obtained from all the databases when updates were made at all the sites in the same domain were consistent 27. ESSIEN, NSEABASI PETER PGS/20002001/4040 79 IMPROVING SOFTWARE CONSTRUCTION PRODUCTIVITY THROUGH REUSABLE COMPONENT INTERGRATION (MAINTENANCE TOOL COMPONENT) 2004 28. OMEIZA, CECILIA OMENEKE PGS/20002001/4040 72 COMPUTER IMPLENTATION OF SOME GRADIENT TECHNIQUES IN NON-LINEAR OPTIMIZATION 2004 (uniform). This shows that the integrity constraint applied is working. Software Development involves designs of computer codes to solve a specific task. It is observed that every software development involving Database Management (DBM) and Management Information System (MIS) software as well must use data maintenance. That is file update codes. From this it is observed that more than 60% of codes written in the development of Software are actually duplicated. The key to solving this problem is in the Software Reuse. It is 'not only the development time that is saved when employing Software Reuse, the quality, measured in term of reliability is equally favored. In this project work codes for file creation, file maintenance (update) and transmission/connectivity is developed christened "MAINTENANCE TOOL COMPONENT". This component serves about 77.0% development time of Database Management Software. That is with this component about 77.0% of codes meant for the development of a standard database software in the area of Payroll, Library System, Material Management Information System are actually reused, not only that, time spent on the Requirement Analysis, Design and Testing stages are also significantly reduced. This Maintenance Tool Component is developed using JAVA programming language. Modeling and Design is based on Unified Software Development Process (USDP) that incorporated Unified Modeling Language (UML). Software in Borland C++ programming language was developed to implement the three popularly used gradient techniques. namely, steepest descent method, conjugate gradient method and variable metric (Quasi-Newton) method for non - linear optimization. Eleven sample problems were selected to test run the software. The performances of these techniques were examined in terms of accuracy, convergence rate, run-time, and 29. MUHAMMAD, PGS/98LAWALI 99/404029 JABAKA NUMERICAL TREATMENT OF SOME WAVE EQUATIONS. 2004 30. ROKO, ABUBAKAR SOFTWARE PACKAGE FOR THE NUMERICAL SOLUTION OF INITIAL VALUE PROBLEMS IN ORDINARY DIFFERENTIAL EQUATIONS. 2004 PGS/9899/404042 storage requirement. It was discovered in general that steepest descent method required the smallest storage size,but relatively slow to converge; conjugate gradient method seems to he the most accurate. Variable metrics was relatively fast to converge but used the biggest storage size and high execution time. Useful recommendations that will assist those who may wish to implement the techniques were made. In this project Finite Element Schemes for solving wave equations were developed. These schemes were implemented using three wave equations, namely, Generalised of d'Alembert wave equation, one dimension hyperbolic equation and telegraph equation for infinite cable. Algorithms were designed and implemented using Borland C++ Programming Language It was observed that the results obtained from the schemes compared favourably with existing analytic solutions especially in the case of generalised d'Alernbert wave equation. In general, the accuracy increases with decreasing mesh size or increase in the number mesh points. Users find it difficult when dealing with implicit RungeKutta methods, due the complexity of iterations and numerical manipulations. In fact the equations defining the stages are usually solved by iterative methods, for instance Gauss-Elimination technique which is very cumbersome. For this reason, this research project is aimed at circumventing this well known difficult through the development of software package. After systematic study of some existing implicit Runge-Kutta methods and Gauss-Elimination/Newton-Raphson methods we designed an algorithm, which was intum, used to develop software package for the simulation of general RungeKutta methods. The software so developed is capable of gererating a large square matrix. Its reliability, accuracy and stability were tested using various initial value problems in ordinary differential equations and the 31. CHIROMA, IBRAHIM NYAM PGS/40400 9 GENETIC ALGORITHM APPROACH TO KNODE SET RELIABILITY OPTIMIZATION FOR A DISTRIBUTED COMPUTING SYSTEM SUBJECT TO CAPACITY CONSTRAINT. 2005 32. ODION, OSHIOKHAIM HELE PHILIP PGS/20002001/4040 49 A MODEL FOR THE OPTIMIZATION OF MILITARY TRAINING (A CASE STUDY OF ARMY WING, NIGERIAN DEFENCE ACADEMY, KADUNA) 2005 results obtained are in fact in good agreement with the exact solutions. One of the important issues in the design of a distributed computing system (DCS) is reliability. The reliability of a DCS depends on the reliability of its communication links and nodes, as well as on the distribution of its resources, such as programs and data files. In the reliability analysis of a DCS, the term K-Node Reliability (KNR) is the probability that all nodes in K (a subset of all processing elements) are connected, where K denotes a subset of a set of processing elements. A K-node set reliability optimization with a capacity constraint is normally a difficult problem and it is evident that it belongs to the class of NP- hard problems. Therefore, heuristic algorithms like genetic algorithms are potential tools for obtaining optimal or near optimal solutions, since they are domain-independent problem-solving approach in which computer programs are evolved to solve or approximately solve the problems. This project research presents an efficient genetic algorithm that computes the reliability of a subset of network nodes of a DCS such that the reliability is maximized and specified capacity constraint satisfied. Compared with existing algorithms (Chen et al, 1994) on various DCS topologies, the proposed GA finds a sub-optimal design much more efficiently. One persistent problem in military training is how to make decisions concerning the proper mix of training aids, conventional training resources, and resource utilization to maintain training proficiency. Given the proficiency standards, the model determines the resources needed and the frequency with which each method needs to be repeated, in order to maintain the standards of proficiency at minimum cost. Data collected from the training of army cadets was modeled using linear programming. The two-phase method was applied 33. FAMUTIMI, RANTIOLA FIDELIS PGS/20002001/4040 58 A COMPARATIVE STUDY OF SEQUENTIAL AND PARALLEL PROGRAMMING IN SOME MATRIX OPERATIONS. 2006 34. SANI, USMAN PGS/9899/404026 A SOFTWARE TOOL FOR PADIATION EXPOSURE MONITORING. A CASE STUDY OF NATIONAL HOSPITALM FOR WOMEN AND CHILDREN ABUJA. 2006 using the Pascal programming language to implement it. The result obtained shows that the training cost reduced drastically by 6.2%. Most of the existing statistical packages used in scientific computations employ sequential processing in their implementation. That is, in solving a particular problem, the sequential program performs some sequential processing in a predetermined order until the whole data have been processed. The sequential program used is regarded as a process. Concurrent processing is a method of processing, whereby the entire processing is further grouped into smaller subgroups and these subgroups are then processed in parallel. Concurrent processing provides a way to organise processing that contains relatively independent parts so as to process them at the same time thereby reducing immensely the overall processing time of the job. It also provides a way to make use of multiple processors. In this way, each independent part (or process) is assigned a processor and all parts are made to communicate so that the individual part's result can be assembled together to obtain the required overall result. This thesis intends to make a comparative study of sequential and parallel programming in some matrix related operations using the response time (speed up) of processing and the memory requirements as parameters. The threat ionizing radiation poses to workers warrants the operation of a personnel monitoring system. In this project report a software tool was developed for a comprehensive film badge services. This tool simplifies the record keeping requirement for maintaining information on individuals who may be exposed to radiation in the cause of their working activities. Its also provides a complete tractability of every potential exposure to radiation that an individual may incur. The system written with visual basic also records personal information (Name, age, sex, address), exposure history 35. ZIRRA, PETER BUBA PGS/20002001/4040 57 SOFTWARE DEVELOPMENT EFFORT ESTIMATION SYSTEM A CASE STUDY OF MICROFOCUS NIGERIA LIMITED, KADUNA. 2006 36. IDRISSA, DJIBO PGS/40401 4 SIMULATION AND DETECTION OF ‘SOFT ERRORS’ THROUGH SOFTWARE FAULT INJECTION AND 2006 and also produces as an output exposure reports of personnel using the film badge. Software effort estimation has posed serious challenges to software developers. Therefore, software practitioners need to understand the effort to be expended and the schedule to be followed in order to make reasonable estimate of resources, to strive to avoid underestimate and overestimate of resources or the amount of time required to deliver the product. Hence, this study identified models of quantitative measures within software product such as the Constructive Cost Model II (COCOMO II) - Post Architecture, Function Points (FPA) and Putnam-Software Life-cycle Models (SLIM) for effort estimation. Primary and secondary sources of data were used to obtain reliable information from the Microfocus Nigeria Limited, Kaduna. A Visual Basic programming, language was used for the implementation of the three models. Results i obtained were analyzed. It was found out that the Microfocus Nigeria Limited uses expert judgment and analogue methods of effort estimation, resulting to poor prediction of staff, time and size. Based on the results of the implementation, some recommendations were proffered to Microfocus Nigeria Limited and other software developers, among others are, that Microfocus Nigeria Limited should adopt the COCOMO II for effort, time and staff estimate, FPA for effective sizing of software project, take into considerations factors affecting effort estimation, and software practitioners should adopt a balanced model for realistic estimate of resources. A systematic approach for introducing data and code redundancy into an existing program written using C++ has been fully described. The transformations aim at making the program able to detect most ofthe soft-errors (bit flip errors) arising in microprocessor-based digital architectures as the consequence of the interaction with Additional work should be carried out towards the definition of a new set of rules, allowing trhe reduction of the resulting overhead (in terms of SOFTWARE FAULT TOLERANCE TECHNIQUES. 37. AJIE, IKECHUKWU JAMES PGS/40401 3 SIMULATION OF SOME MULTIPLE CRITERIA DECISION MAKING PROBLEMS USING FUZZY SET APPROACH. 2006 38. YUSUF, MOHAMMED AGETEGBA PGS/20012002/4040 98 DATA SECURITY: LAYERED APPROACH ALGORITHM 2007 radiation. The fault coverage obtained by the method, as well as the slow-down and code size increase it causes, were investigated. for the proposed transformation used on all the selected benchmarks (bubble sort, matrix multiplication, and Factorial routines), it has been Found that the average increase in the source code size (in terms of number of lines) was 1.82, the average i ncrease ill the executable code size (in terms ofnumber of bytes) was 2.3 and the performance slow-down introduced by the application of the rules was about 2.66. Human beings have always been faced with the problem of data processing and decision making based on the data. This problem increases when it involves choosing between alternatives based on multiple criteria. This work used fuzzy set approach to study decision making under multiple criteria. The level of commitment of members of a social club and also the grading system or students was studied. The resultant models were implemented using C++ programming language. The concept of fuzzy similarity degree was used to dcicrm i nc total scores of a set of students. The result of the simulation showed that it is very efficient in breaking ties experienced when the normal method based on percentage is used. The notion of membership function was also used to arrive at a model that was used to determine the level of commitment of members of a social club. Two different weighting methods were used to implement the model. The results of the implementations showed that the decisions taken were consistent with one another. It also showed that the level of commitment of members could be determined quantitatively. In this study we adopted the method of Two-FactorAuthentication strategy. Two-Factor-Authentication is a security process that confirms user identities using two distinctive factors, that is something they have (this memory and speed) at the cost of slightly reduced fault coverage capacity. More research should also be carried out to complete the set of rules to cover all the costructs of a high-level language. Further research work is recommended in the implementation of the encryption program with 39. ADEGOKE, GBOLAGADE KOLA PGS/20012002/4040 119 COMPUTER SIMULATION PRODUCTION AND OPERATIONS MANAGEMENT (A CASE STUDY OF DEBO INDUSTRIES LIMITED, LAGOS). 2008 40. ABBA, SANI PGS/20052006/4040 SIMULATION OF PERSISTENT 2008 factor includes keys, cards, tokens and so on) and something they know (such as Password and Pins), except that in this case we will not use physical objects like cards, token and physical keys, only encryption keys and password that will be applied. By requiring individuals to supply two different credentials will reduce the risk of fraud and make our data or information more secured. The Two-Factor-Authentication method employed by this study combined Trigger-ChallengeResponse password algorithm and Data Encryption Standard algorithm to provide double protection. The software design thus, has provided two layers of protection by authenticating individuals twice, there by offering much resistance to single-factor attacks, because user has to know both Key and Password. This project report concerns simulation of production and operations management as an essential activity in a manufacturing sector. Production and operations management is currently one of the most research areas in the manufacturing sector of the economy. The reason is that production and operations management facilitates the effective and efficient accomplishment of production objectives of an organization. Some mathematical and statistical models were used to compute the expected process duration, actual process duration, economic order quantity whose purpose is to reduce costs, and machine breakdown frequency whose purpose is to predict machine breakdown and for planning and preventive maintenance. Test data were gotten from Debo Industries Limited, Lagos and were simulated. The results generated were used to predict a real life situation and have been presented and discussed. The language of implementation of the model is Turbo Pascal due to its capability as a scientific programming language. This research work is an attempt to design a simulated prototype to model the actual kernel operation of other encryption algorithm such as Triple DES (TDES) or other such algorithm that will provide significantly higher level of security. Further work to be carried out on distributed Nano- 41. ALI, ISA IBRAHIM 371 NANO-BASED SECURITY OPERATING SYSTEM KERNEL, IN WINDOWS XP ENVIRONMENT. PGS/20052006/4040 375 ISLAMIC INHERITANCE SYSTEM USING OBJECT ORIENTED DEVELOPMENT 2008 Windows XP operating system and to evaluate the Kernel for fault containment performance based on micro benchmark and kernel in distributed environment. profiling performance metrics. A new prototype namely, Persistent ano-Based Security Operating System Kernel (P SOSK) was designed and implemented using Microsoft Visual C++; Application Programming Interface (API), and Unified Modeling Language (UML). The simulated kernel competes with the Windows XP kernel functions such as: handling hardware exceptions and interrupts; scheduling; prioritizing; dispatching threads; synchronizing execution across processors in a multiprocessor environment; kernel security monitoring; system initialization and startup (Windows Registry). The prototype ran on an Intel IA-64 hardware platform (Intel Pentium IV MMX Dual core 2.8GHZ processor) Computer, running Microsoft windows XP professional version 2007 Service Park 3 operating systemnumerical experiments on random instances of the performance metrics suggest that the P NSOSK model performs satisfactorily in windows XP's environment. The mutual exclusion primitives, synchronization and concurrency problems were solved using the model. Islamic inheritance system using object oriented development was focused with the view of producing a computerized inheritance system. However, inheritance is a very complicated issue in Islam, and as well it is a serious trial and affliction on Muslims. Sharing justly is not just a necessity but also an obligation if and only if the judge wants to succeed in this world and even in the hereafter. Quite a number of verses and few prophetic traditions encourage Muslims to get hold on to this knowledge and remind them the risk of loosing it. Also investigated were, factors concerning risks loss when Muslims deal with inheritance in accordance with their own selfish interest and not in accordance with the 42. ALIYU, SULEIMAN ONIMISI PGS/200506/404037 0 DESIGNING AN INTERACTIVE ENVIRONMENT FOR LEARNING FRACTIONS IN PRIMARY MATHEMATICS USING OBJECT ORIENTED SOFTWARE ENGINEERING 2009. rulings set out by their creator. Procedures for allocating and allotting right fraction/ratio to each and every heir/inheritor receives legal proportion based on the verdict of our pious predecessors were developed according to the Islamic injunction. Unified modeling language was selected for the development, flow chart for the simple design and Java programming language was selected for the implementation. Database has been created in order to assist in solving the problem manually and it has been simulated for the implementation. The output obtained, really achieved the desired goals which has been shown in the result. With the tremendous growth of the use of computers in schools, sound research is needed on how to design interactive learning environments that effectively help children by promoting reflective cognition and better learning. The research described in this thesis addresses the following issues in designing interactive mathematics learning environments for children: a) The user interrace design should support children's learning of mathematical concepts. b) Design features should be effective in promoting reflective cognition and better learning. c) The learning environment should be tailored to meet children's affective needs. Psmath-fraction is an interactive learning activity aimed at assisting elementary school children in understanding fractions was developed. Techniques such as visual feedback and scaffolding were used in the design to promote reflective cognition. The result of this research provides a motivating learning environment and the features of the software matched children's interests and are conducive to children's enjoyment of the learning activity. The current research suggests some directions for future research: 1. Research is needed to develop guidelines and principles on how to design effective navigational structures and sequencing of activities for different domains. 2. Research is needed to develop guidelines and principles on how to design different styles of interfaces suitable for different types of learners. 3. Research is needed to investigate how to precisely assess users' understanding of the domain in order to provide more targeted and adaptive feedback. 4. Research is needed to investigate what types of entertainment elements promote children's motivation without distracting them from the exploration of the mathematical concepts. 43. YAHAYA, MOHAMMED ONIMISI PGS/0506/404036 6 DEVELOPMENT OF SEQUENCE AND SERIES ANALYSER SOFTWARE: AN INTEGRATED SOFTWARE FOR MATHEMATICAL SEQUENCES AND SERIES ANALYSIS 2009 44. UMAR, FATIMA ZAMBUK PGS/20052006/4040 374 COMPARATIVE STUDIES OF SOFTWARE ARCHITECTURES 2009 45. YA’U ABDULSALA PGS/20052006/4040 DESIGN AND IMPLEMENTATION 2009 In this work, language design principles in computer science and notions related to sequences in mathematics were studied. The evolutionary prototyping design method is used to develop a software called Sequence and Series Analyser. It extended the features of Sequencer and Series plotter by including features like sequence generation, sequence analysis, special sequences like Fibonacci sequence, sequences of partial sums and series, sequence prediction and mathematical induction. The field of Information System is experiencing major changes as the underlying technology moves from centralized, host based computing to distributed, client server computing. Computer transaction using client server model are very common. Client server model provides a convenient way to interconnect and gather information that is distributed at different area. In this thesis, a study of a number of common software architecture styles upon which many systems are currently based on was made. A web base application software designed using software methodology was used to measure the effectiveness, efficiency, maintainability, re- usability and propagation delay-time performance of Repository and two tier Client server Architecture. Client server was found to be more efficient, effective, it has high rate of component reusability and high maintainability compared to repository. The Mathematical Sciences Programme of Abubakar Tarawa Balewa University, Bauchi, has been using M GITAL 362 OF INTERNETBASED INFORMATION SYSTEM FOR MOBILE DEVICES ( CASE STUDY OF MATHEMATICAL SCIENCES PROGRAMME ATBU BAUCHI) 46. SALAMI, HAMZA ONORUOIZA PGS/0506/404036 3 AN INTERGRATED ASSEMBLY LANGUAGE SIMULATOR (PENSIM) 2009 47. SAMBO, AMINA SANI PGS/20012002/4040 110 COMPARATIVE ANAYSIS OF INTERNET BROWSERS 2010 papers and ink since the inception or the University to pass information to both Staff and Students of the department. In this project, an Internet Based application called Math _info for hand held devices was designed to support the activities of mathematical Sciences Programme. Basic activities such as viewing student results, Lecture time table and the courses registered by individual student in the programme were considered for implementation in this application. The application also contained the Profile of all Academic staff in the programme. Staff can also access courses allocation for a particular Semester. A study of existing microprocessor simulators was made, and their limitations were identified. These simulators do not support 32- bit registers and instructions. The object oriented analysis and design methodology was then used to develop PENSIM (the Pentium Simulator). The simulator was tested and found to be very reliable. It extended some of the features of existing microprocessor simulators such as number of registers and instructions implemented. PENSIM also provided a more detailed tutorial compared to those found in the existing simulators. As more and more services become available on the Internet, the issue of fast and secured access to online resources gains more importance. The growth of the internet has encouraged a high number of people to explore and take advantage of the World Wide Web (www). The window to the World Wide Web is a web browser, hence the development of various web browsers in the market today. A comparative study of four web browsers namely netscape navigator, mozilla firefox, internet explorer and Opera was carried out. A web portal which was fully developed and tested was used to evaluate the performance of the four browsers. The results revealed that mozilla fire fox and netscape Based on the research, I recommend that users should have an idea about the kind of site they want to visit and based on that they should use a specific browser to optimize their browsing. 48. CHAKA, JOHN GYANG PGS/20012002/4040 126 COMPARATIVE STUDIES OF ANTIVIRUS SOFTWARES IN NETWORK SECURITY 2011 49. TELLA, MUHAMMAD PGS/20012002/4040 91 SIMULATION OF A GROUP DECISION MAKING (GDM) MODEL FOR MANAGEMENT DECISION SUPPORT 2012 perform best in download time and page layout, Internet explorer is best in conserving memory usuage, privacy and security. Opera performs best in speed and performance. This thesis compares the performance of four (viz: AVG Antivirus 9.0, Syrnantec Norton Antivirus 2010, Kaspersky Internet Security 2010, and Avast Antivirus 5.0) of the most widely used antivirus software in order to ascertain the most efficient and effective in handling network threats. A comparison of the performances of these antivirus software is needed to enable users make informed choices in the use of any of the antivirus software. AVG Antivirus 9.0 was found to be more effective, efficient and has a high performance compared to Syrnantec Norton Antivirus 2010, Kaspersky Internet Security 2010, and Avast Antivirus 5.0. Group Decision Making (GDM) is one of the approaches that have gained wide acceptance in organisations. In GDM problems, a set of experts are faced with the choice of the best alternative from a set of many alternatives. A GDM model with incomplete Fuzzy Preference Relations was developed to overcome the problem of inconsistency in incomplete fuzzy preferences. In this study, six laptop computers with different specifications were used for the simulation. It is expected that given any set of alternatives, the software should be able to monitor the consistency or inconsistency of an expert in giving his preference value for the pair wise comparisons of the alternatives and be able to suggest alternative preference values for any given pair for which the expert cannot provide a preference value. The research investigated the reliability of the model when the number of alternatives is too large to cope. The work is designed along two different aspects: Interface Requirements, and Logical goals with use of Java as the programming language in the coding of the models. Further research work can be carried out simulate several seenarios in which many decision makers are involved. 50. BELLO, MUHAMMAD ALIYU PGS/20002001/4040 69 SOFTWARE TOOL FOR IMPLEMENTATION OF GAUSSLOBATTO RUNGEKUTTA METHODS 2012 51. MAISHANU, MARYAM PGS/20012002/4040 89 AN INTERFACE FOR A PC BASED OSCILLOSCOPE 2011 The increasing gap between the speeds of processors and main memory of a computer has led to hardware architectures with an increasing number of caches to reduce average memory access times. Such deep memory hierarchies make the sequential and parallel efficiency of computer programs strongly dependent on their memory access pattern. In this thesis, we consider implicit GaussLobatto-Runge-Kutta collocation methods for the solution of ordinary differential equations and study their efficient implementation on different parallel platforms. In particular we focus on systems of ordinary differential equations which are characterized by large Lipschitz constants. This class of equations is usually very stable but often difficult to solve even numerically because of their fast responding component which tends to control the stability of numerical methods. Here we explore how the potential parallelism in the stage vector computation of such equations can be exploited in a pipelining approachleading to a better locality behavior and higher scalability. Experiments show that this approach results in efficiency improvements on several recent sequential and parallel computers. A PC based oscilloscope was developed with appropriate interface to an electronic device specifically the PC sound card with its integrated ADC using MATLAB data acquisition tool. The interface developed is easy to understand and flexible enough to handle data acquisition tasks with various configurations. The system was tested and the results of sampling with different trigger settings were observed by adjusting duration, timeout, trigger repeat and threshold values for the advanced trigger. Data exploration tools were used to explore data in various ways after the signal has been captured. Other features of the system tested include data export to an excel file, generation of subset of the data using suplot panel and how exported image could be displayed in a graphics Further research in this area should consider integrating the MATLAB Signal Processing Toolbox to improve signal analysis and measurement capabilities for the software. 52. UMOH, NNAMSO MICHAEL PGS/0607/404039 6 SOFTWARE RELIABILITY LEVELS CASE STUDIES OF SOME SOFTWARE COMPANIES. 2013 53. MOHAMMED, DANLAMI PGS/20062007/4040 397 QUERY OPTIMISATION AND PERFORMANCE ANALYSIS ON WEB 2013 application. When put into use, the software could contribute to learning and self development as users gain insight into how MATLAB and DAQ toolbox are utilized in developing applications that meet the requirements of more specific data acquisition tasks. Software reliability is a major factor in software quality and is the index by which users judge a particular software. Software reliability to the users is the ability of the software to perform its perceived function without fail. It therefore must be determined in all software projects of importance. Some software reliability models were reviewed and the Jelinski-Moranda model was chosen which was considered the most appropriate for the intended software tool (i.e Re1Soft). Some software houses were visited and failure data collected, formatted and analyzed. The least square method was used to determine some parameters such as Et (i.e. total error) and K, the proportionality constant. These parameters were used as input to design Relsoft tool which was then used in determining the reliability levels of the selected software produced by Nigerian software companies. The result obtained from RelSoft, showed that nine out of the ten selected software have reliability levels ranging from 0.507 to 0.825, except one which has a reliability level of 0.425. Other reliability indicators such as failure rate and MTTF (i.e. Mean Time To Failure) were also calculated using RelSoft. The software reliability is measured on a scale ranging from zero to one (i.e. 0 to 1). However, for life-critical application the reliability must be in the region of 0.999 to 1.0 for it to be considered to be reliable. Query Optimization is the process of choosing the most efficient way to execute a Structured Query Language (SQL) statement and web services is a software system designed to support interoperable machine-to-machine interaction over a network. These two technologies when Further research on this field should also consider trying other programming language for the implementation such PHP and MySQL, Java SERVER integrated offer better performance of application by eliminating machine-to-machine dependence. Using sample sizes of 1000, 2000 and 3000 records, the thesis uncovered that adding more memory to the system improves the performance of the Relational Database Management Systems (RDBMS). In another case, changing Hard Disk Size has little effect on the performance of RDBMS. Also, adding more data to the database has effect on the performance of the RDBMS. While changing the processor size from single core to duo and quad core improve the performance of the RDBMS. In all the cases stated above, SQL Server 2005 offer better performance followed by Microsoft Access 2003, Oracle 10g and My SQL 5.1 respectively. The application is recommended for RDBMS developers. Programming Language. It is recommended that future research on this area should consider increasing the sampling size. This is to visibly eneble the researcher see the effect of the hardware change clearly.