Multi objective mixed model sequencing problem with just in time assembly line and an application

MULTI OBJECTIVE MIXED MODEL SEQUENCING PROBLEM WITHJUST IN TIME ASSEMBLY LINE AND AN APPLICATION(M.Sc. Thesis)Salih Hakan AKÇINGAZ UNIVERSITYINSTITUTE OF SCIENCE AND TECHNOLOGYSeptember 2006ABSTRACTIn this study, the sequencing problem in an just in time assembly linewith multiple objectives examined. The objectives that are taken in toaccount are minimization of number of setups and minimization ofusage rate. In the literature survey related approaches for the problemexamined. To solve the problem ant colony heuristic which is a meta-heuristic approach is applied.To apply the ant colony heuristic approach for this NP-Hard typeproblem with the objective of minimization of usage rate and number ofsetups , travelling salesman problem technique is used. Byspatialization of the problem, with the help of travelling salesmanproblem, solutions are generalized. At the phase of problemspatialization , for different weights of number of setups and usagerates, six diffrenet strategies (KKO1, KKO2, KKO3, KKO4, KKO5, KKO6)generated.Also by improving these six strategies by changing the usage ratecalculation method, six new strategies generalized(G-KKO1, G-KKO2,G-KKO3, G-KKO4, G-KKO5, G-KKO6). These six improved strategiesand random sequencing strategies are also examined.Artificial intelligence algorithms of ant colony heuristics are coded byDelphi programming language. The performances of the strategies areidentified by an application on an industrial problem data. It is appliedfor a just in time mixed model sequencing problem for an assembly linewhich produces three different products with a total amount of 180pieces in a 450 minutes shift. Generated sequences examined by CPUtime and usage rate performance metrics. For the performance metricsit is observed that KKO1, KKO2, G-KKO1 and G-KKO2 strategiesoutperforms other strategies.: 906.1.141Science CodeKey Words :aSequencing problems, heuristic methods,aaaaaaaaaaaaaaaaaassembly line, just in timePage Number : 51Adviser : Assoc. Prof. Dr. Ertan GÜNER