Adjacency – Based Heuristic for Facilities Design Ornurai Sangsawang and Sunarin Chanta Department of Industrial Management Faculty of Technology and Industrial Management King Mongkut’s Institute of Technology North Bangkok 1518 Pibulsongkram Rd., Bangsue, Bangkok 10800, Thailand Phone / Fax: 6637 213 367 Ext. 7042 Email: oss@kmitnb.ac.th Abstract This article presents a new construction algorithm for computer – aided plant layout. A case study is an assembly plant which will be expanded to construct a new plant. The layouts are generated by ALDEP and Adjacency – Based Heuristic that are evaluated by the maximize adjacency – based objective. The solution will be developed on the basis of mathematical expressions that can be evaluated objectively. For the results, found that the layout generated from Adjacency – Based Heuristic has the layout score better than the layout is generated by ALDEP, and Adjacency – Based Heuristic can generate the material handling distance less than the layout is generated by ALDEP. 2. Method Maximize adjacency – based objective. That considers the adjacency score is computed from the sum of all relationship values which considers the adjacency departments only. m = the number of departments wrij = weight for the adjacency class, corresponding to the relationship chart (A = 64 , E = 16 , I = 4 , O = 1 , U = 0) daij = adjacency distance from department i to department j daij = 1; department i and department j are adjacency daij = 0 ; department i and department j are not adjacency Keywords : Computer – aided plant layout , ALDEP, Adjacency – Based Heuristic, construction algorithm, Facilities Design. 1. Introduction Modern industries have rapidly developed manufacturing system. The layout design is an important task when a manufacturing system is constructed, or expanded. If the facilities are arranged optimally, manufacturers can decrease work – in – process, material-handling costs, total production costs and significantly enhance a system’s efficiency. Computer – aided plant layout has several advantages. The computer can perform the calculations and generate several solutions much more accurately than manual procedures and the solutions will be developed on the basis of mathematical expressions that can be evaluated objectively. Facilities design has the major objective of cost minimization, and the material handling cost can be reduced by placement closely related facilities. A good layout leads to reducing production costs, and increasing productivity. Construction layout algorithm is the important procedure to generate a layout in the first time. There are varieties of selection and placement procedures in several construction algorithms. Each algorithm has the suitable objective for itself such as minimize distance – weighted adjacency – based objective , maximize adjacency – based objective to maximize closeness. 2.1 Selection procedure The first department is selected to enter the layout from the department having the lowest Total Closeness Rating (TCR). The second department is selected from the department having strongest relationship with the first selected department to be a victor. Then, select the next department from the department that has the highest important relationship (A or E) with the victor. If no departments have the important relationship with the victor, the next department to enter the layout will be selected by consideration the highest sum of relationship between the department and all the selected departments. This process continues until all departments have been selected to enter the layout. From the relationship chart in figure 1, department 1 has the lowest TCR, it is the first to enter. The next department, which has the strongest relationship with the first selected department, is department 2.Then, search for the highest important relationship with department 2, and there is not found. Later, try to find the highest sum of relationship among all the unselected departments and the unselected departments. Found that department 3 has the highest sum of relationship with department 1 and department 2. Thus, the next victor is department 3. Then, Department 4 is selected from the highest relationship with department 3 to be the next victor. Selection Proceedings of the International Conference on Computer and Industrial Management, ICIM, October 29-30, 2005, Bangkok, Thailand 15.1 of the next department depends on the highest important relationship that it has with the previously selected department. The overall order of departments entering the layout is 1-2-3-48-9-10-7-6-5. Figure 3 the layout generated from ALDEP algorithm Table 1 layout score computation of the layout from ALDEP algorithm 2.2 Placement procedure Placement procedure begins by placing the first department which derived from selection procedure in the right-downside corner of the layout. And placing the next departments uses the placing rating that is the sum of the weighted closeness ratings (A=243, E = 81, I = 27, O = 9, U = 1, X = -729) between the entering department and its new neighbors. The next departments are entered at that location with the largest placing rating. Adjacency Department 1-2 1-3 1-4 Relationship E I U Score 16 4 0 1-9 2-10 3-4 3-9 5-6 5-8 U U E U A U 0 0 16 0 64 0 6-7 6-8 7-8 7-10 8-9 9-10 A I A U A A 64 4 64 0 64 64 Total score Figure 1 the relationship chart for electronic printed circuit board As a result, the layout is generated by Adjacency Based Heuristic, which is displayed in the figure 4. The ordinal selected departments are 1-2-3-4-8-9-10-7-6-5. The layout score is 373. 3. Results In the case study created a new layout for electronic printed circuit board by using ALDEP and Adjacency – Based Heuristic. The plant layout was generated by ALDEP which the ordinal selected departments are 5-6-7-8-9-10-2-1-3-4. By referencing maximize adjacency – based objective. The layout score is equal to 360. the minimize distance – weighted adjacency – based objective, the layout is displayed in the figure 3. 5 5 8 8 9 9 3 3 4 4 5 5 8 8 9 9 3 3 4 4 5 5 8 8 9 9 1 1 4 4 5 5 8 8 9 9 1 1 4 4 6 6 8 8 9 9 1 1 6 6 8 8 9 9 1 1 6 6 7 7 10 10 2 2 6 6 7 7 10 10 2 2 360 Special Issue of the International Journal of the Computer, the Internet and Management, Vol. 13 No.SP2, October, 2005 15.2 Start Read the input data - Relationship chart - each department areas Define the selected dep. as the winner Select the department having the lowest TCR department Select the department having the strongest Rel. with the winner Are the dep. have the highest‘Important’ Rel. with the victor ? Define the selected dep. as the victor y Select the dep. having the highest‘Important’ Rel. with the victor n Are the dep. have the highest ‘Important’ Rel. with all the selected deps. ? Define the next victor n y Select the dep. having the highest‘Important’ Rel. with all the selected deps. Are all the dep. already selected ? Figure 2 Flowchart for Computer Aided Layout Planning with Department Relationship y Placement Procedure n Calculate the sum of the Rel. with all the selected departments Display the layout Compute the layout score Select the highest sum of the Rel. with all the selected departments End Proceedings of the International Conference on Computer and Industrial Management, ICIM, October 29-30, 2005, Bangkok, Thailand 15.3 5 5 7 7 8 8 4 4 4 4 5 5 7 7 8 8 4 4 4 4 place at the right – downside corner of the layout. Similarity, ALDEP focuses on the strongest relationship with the immediately previously selected department, and so does Adjacency-Based Heuristic. If no departments have the important relationship with the selected department, ALDEP procedure will randomly select the next department while Adjacency-Based Heuristic will be selected by consideration the highest sum of relationship between the department and all the selected departments. 5 5 6 6 8 8 3 3 2 2 5. Suggestion 5 5 6 6 8 8 3 3 2 2 6 6 8 8 1 1 1 1 6 6 8 8 1 1 1 1 5.1 Adjacency-Based Heuristic is a construction algorithm that can reasonably generate a good layout, but it cannot confirm that the layout which found is the optimal layout. 5.2 Adjacency-Based Heuristic operates faster than ALDEP algorithm, which randomly generate several times and many layouts until the good layout is constructed. 5.3 The reality objectives for facilities design such as flexibility and safety should be considered. 10 10 9 9 9 9 9 9 10 10 9 9 9 9 9 9 Figure 4 the layout generated from the Adjacency – Based Heuristic Table 2 layout score computation of the layout from Adjacency – Based Heuristic 6. References D.R. Sule (1998), Manufacturing Facilities Location, Planning, and Design: International Thomson Publishing Adjacency Department 1-2 1-3 1-8 2-3 Relationship E I U O Score 16 4 0 1 2-4 3-4 3-8 4-9 4-8 5-6 U E U U E A 0 16 0 0 16 64 5-7 6-7 6-8 7-8 8-9 9-10 U A U A A A 0 64 0 64 64 64 Total score Edwards, H.K., Gillett., and Hale, M.C. (1970). Modular Allocation Technique; MAT. Management Science, Vol. 17, 161-169. Heragu Sunderesh (1997). Facilities design. Imprint Boston: PWS Publishing Company. Tompkins, James A., and James M. Moore (1977). Computer Aided Layout: A User’s Guide. American Institute of Industrial Engineers, Norcross, GA. 373 4. Conclusion From the results as shown in the table 1 and the table 2, found that the layout generated from Adjacency – Based Heuristic has the layout score better than the layout generated from ALDEP. In the selection procedure of Adjacency – Based Heuristic, to reduce the opportunity which the edge area of the lowest TCR department will be adjoined with the other departments. The lowest TCR department is firstly selected to Special Issue of the International Journal of the Computer, the Internet and Management, Vol. 13 No.SP2, October, 2005 15.4