BIOGEOGRAPHY-BASED OPTIMIZATION (BBO) ALGORITHM: A REVIEW Manjeet Kaur*, Sakshi Sethi Assistant Professor, Amity University, Gurgaon, Haryana, India Abstract - In this article we are presenting review of an optimization technique biogeography based optimization algorithm. Although there are several optimization techniques like genetic algorithm, ant colony, and more yet BBO algorithm is meta-heuristics search optimization technique largely focus on overall optimization. Here we review various methods under BBO algorithm for optimization. Keywords- Optimization, BBO, Operators. I. Introduction: Optimization is the act of achieving the best possible result under various given circumstances. The primary aim of the optimization techniques is either to minimize effort or to maximize benefit. The effort or the benefit can be usually expressed as a function of certain design variables. Hence, optimization is the process of finding the conditions that give the maximum or the minimum value of a function. In this paper we are trying to understand how biogeography based optimization algorithm provides solution for optimization problem. BBO is a part of Evolutionary Algorithms that deals with natural selection in the search space and selects the best solution. The BBO works for global optimization. Hence this paper represents review of BBO algorithm for optimization. II. Types of Optimization Techniques Optimization problems are set of problems that find the best result either by minimizing the cost or maximizing the fitness. There are wide ranges of application where optimization is needed for speed and position control of electric motors, travelling salesman problem, root finding for system of linear and nonlinear equations, tuning the controller parameters, circuit and other. In general, what we have to do is to reduce the cost or increase the profit. Different Optimization techniques are: Genetic Algorithm: Genetic algorithm is a heuristic optimization technique inspired by the mechanisms of natural selection. It starts with an initial population containing a number of chromosomes where each one represents a solution of the problem in which its performance is evaluated based on a fitness function. Based on the fitness of each individual and defined probability, a group of chromosomes is selected to undergo three common stages: crossover and mutation [10]. The application of these three basic operations is to allow the creation of new individuals to yield better solutions then the parents, leading to the optimal solution. Ant Colony: Ant colony method simulates the behavior of ants living in colony. Initially all ants start searching for food in arbitrary manner. Once a path is found by one ant which leads to food all other will follow that path. BBO: Biogeography is the study of the geographic distribution of biological organisms. In BBO model, problem solutions are represented as islands and the sharing features between solutions are represented as immigration and emigration between the islands. III. Basic Concept of BBO Mathematical models of biogeography describe migration, speciation, and extinction. Islands that are well suited as habitats for biological species are said to have a high habitat suitability index (HSI). Features that correspond with HSI include topographic diversity, area, temperature, rainfall, etc. and the variables which specify these characteristics [2] are called suitability index variables (SIVs). SIVs are independent variables of the island, and HSI’s are dependent variables. Islands with high HSI tend to have a large number of species, and those with low HSI have a small number of species [3]. Islands with a high HSI have many species that emigrate to nearby islands because of the accumulation of random effects on its large populations. Emigration appears like animals ride, flotsam fly or swim to other neighboring islands. Biogeography-based optimization (BBO) algorithm was first presented in [2] and is an example of how a natural process can be modeled to solve optimization problems. Biogeography is the nature’s way of scattering species, and is similar to general problem solutions. A good solution is homologous to an island with a high HSI, and a poor solution is just like an island with a low HSI. High HSI solutions have more capability to share their features with other solutions, and low HSI are more likely to share their features from other solutions. This criterion to problem solving is called biogeography-based optimization (BBO). Fig.1 illustrates a model of species abundance in an isolated island. The emigration rate µ out of the island, and the immigration rate λ into the island, are functions of the number of species S on the island. The maximum possible immigration rate I occurs when there are zero species on the island. As the number of species expands, the islands become more crowded, fewer species are able to successfully survive immigration, and the immigration rate decreases. The largest possible number of species that the island can reinforce is Smax, at which point immigration rate is zero. If there are no species on the island then the emigration rate is zero. As the number of species increases, the islands become more crowded, representative individuals of species are more likely to leave the island, and the emigration rate increases. The maximum emigration rate E occurs when the island contains the largest number of species that it can support. Emigration Rate E or I Immigration Smax No. of species Fig. 1 Distribution of Species in an Island We have shown the immigration and emigration curves in fig. 1 as straight lines, but in general they might be nonlinear. This model simply gives a general description of the process of immigration and emigration. IV. Biogeography-based optimization algorithm: In a group of neighboring islands, species of plants and animals will migrate over time between the islands by various means, being carried along by driftwood [2], fish, birds, and the wind. Over evolutionary period of time, some islands may tend to accumulate more species than others because they possess certain environmental features that are more suitable to sustaining those species than islands with fewer species. This ability to sustain large number of species can be associated with a fitness measure that we can quantify by assigning a habitat suitability index (HSI) to each island. The value of the HSI depends on many features of islands. If a value assigned to each feature, then the HSI is the function of each feature. Each of these values is represented by suitability index variable (SIV). These mappings are summarized as follows: Island (feature1, feature2,…, featuren) (SIV1 SIV2,…,SIVn) HSI An island with a large number of species (a high HSI) has an abundance of species which can immigrate to other islands, so its rate of emigration, denoted by µ, is correspondingly large. The island is also less likely to able to sustain further immigration of species because of growing demand on its finite environmental resources, so its immigration rate, denoted by λ, is small. For many applications it suffices to assume a linear relationship between an island’s HSI and its immigration and emigration rates and that these rates are the same for all islands under consideration. These relationships are depicted in fig. 2. One can use the migration rates of each solution to probabilistically share features between islands. For each SIV in each island, it is probabilistically decided whether or not to immigrate. Emigration Rate Immigration E or I S1 S2 Fitnes E or Fig.2. Illustration of two candidate solutions to s some problem. S1 is poor solution and S2 is good solution. I If immigration is [3] selected for a given SIV, then the emigrating island is selected probabilistically using Roulette wheel selection normalized by µ. After the migration operation, a mutation operation is probabilistically applied to the island to increase diversity in the population. This gives the algorithm shown in fig. 3 as [4] a conceptual description of one generation using this approach (Single Immigration BBO Algorithm). We use the notation yk (s) to denote the sth feature of the kth island in a population y of islands. Migration and mutation of the entire population take place before any of the solutions are replaced in the population, which requires the use of the temporary population Z [3]. Z Y For each island Zk For each SIV s Use λk to decide whether to immigrate to Zk(s) If immigrating ting to Zk (s) then Use µ to select the emigrating island yj Zk(s) yj (s) End if Probabilistically decide whether to mutate Next SIV s+1 Next island Zk+1 Y Z Fig. 3 One generation of the BBO Algorithm As with other population-based algorithms, we often incorporate elitism in order to retain the best solutions in the population from one generation to the next. We use Z to denote the no. of top individuals in the population that have a zero probability of immigration [4]. If elitism is not used, then Z = 0 and immigration rate λ is a linear function of fitness and is positive for all fitness values. If elitism is used, then Z > 0 and the immigration probability is zero for the top Z individuals in the search space. fmin fmax fmin fmax fmax a) Immigration rate with no elitism (Z = 0). Immigration rate is a function of fitness. Table 1: Parameters used in BBO Parameters of BBO Probability of modification Probability of mutation Keep (Elitism parameter) Values 1 0.005 2 VI. CONCLUSIONS & FUTURE WORK λ λ V. Review of Selection of BBO Parameters: To start up with BBO, certain parameters need to be defined. Selection of these parameters decides to a great extent the ability of global minimization. The migration operator affects the ability of escaping from local optimization and refining global optimization. Mutation operator is used to obtain required diversity among the population. Elitsm operator retains the best solutions among the population. Initializing the values of the parameters is as per table.1. b)Immigration rate with elitism (Z ≠ 0). The immigration rate λ is zero for most fit Z individuals. Fig. 4 Plot of fitness vs. immigration probability (a) Without elitism, (b) With elitism We have successfully reviewed the BBO algorithm for the optimization problem. It is clearly understood from this paper how BBO algorithm originated, the basic concepts and the working of BBO algorithm. BBO algorithm is one of the well-known optimization techniques. Other technique like genetic algorithm (GA), ant colony optimization and particle swarm optimization (PSO) are explained. In future we will design modified version of BBO algorithm and run, test it for various optimization problem. Also we will compare BBO with other optimization technique. REFERENCES [1] Dan Simon, “Population Distribution in Biogeography-Based Optimization Algorithms with Elitism,” IEEE International Conference on systems, man, & Cybernetics San Antonio, USA-October 2009. [2] Dan Simon, “Biogeography-Based Optimization,” IEEE transactions on Evolutionary Computation, vol.12, no. 6, PP. 702-713, December 2008. [3] Dan Simon, “Markov Models for BiogeographyBased Optimization,” IEEE International Conference on systems, man, & Cybernetics San Antonio, USA, 2010. [4] Dan Simon, “Population Distribution in Biogeography-Based Optimization using Elitism,” IEEE International Conference on systems, man, & Cybernetics San Antonio, USA, 2009. [5] Dan Simon, “A Probabilistic Approach of Biogeography-Based Optimization Algorithm,” IEEE transactions on Evolutionary Computation, vol.12, no. 6, PP. 1009-1014, February 13, 2009. [6] Dan Simon, “Oppositional Biogeography-Based Optimization,” IEEE transactions on Evolutionary Computation, vol.12, no. 6, PP. 702-713, October 2009. [7] Dan Simon, “Biogeography-Based Optimization & the solution of the Power Flow Problem,” IEEE International Conference on systems, man, & Cybernetics San Antonio, USA-October 2009. [8] Dan Simon, “Biogeography-Based Optimization for Robot Controller tuning,” IEEE International Conference on systems, man, & Cybernetics San Antonio, USA, 2009. [9] S. Kumar, et.al, “Fuzzy Rule Base Generation from Numerical Data using Biogeography Based Optimization,” Computational Intelligence Laboratory, Institute of Science and Technology Klawad, PP. 1-6, 2009. [10] A.H. Wright, “Genetic Algorithms for real parameter optimization: Foundations of Genetic Algorithms” J. E. Rawlins, Ed. San Mateo, CA: Morgan Kaufmdnn, 1991, pp. 205-218. [11] K.F. Mann, Member, IEEE, K.S. Tang, S. Kwong, Member, IEEE, “Genetic Algorithms: Concepts and Applications”, pp 519-534. [12] C.A. CoelloCoello, G.B. Lamont and D.A.V. Veldhuizen, “Evolutionary Algorithms for Solving Multiobjective Problems” 2nd ed. Springer. [13] Saurabh Kumar, “Multi Objective Ant Algorithm to Parameter optimized of PID Controller and Compare to Genetic Algorithm”, Computational Intelligence and Computing Research (ICCIC), 2013, pp. 1-4. [14] J.S. Saini et. al, “Genetic Algorithm Based PID Tuner,” journal of institution of Engineers (India) Electrical Engg. Divn, vol. 85, pp. 216-221, Mar. 2005.