Optimal Sequence Identification in Parallel Flow Line Environment Using Heuristics

In manufacturing environment, it is necessary to find the sequence of jobs on different machines; so that technological constraints are satisfied and the performance criteria are optimized. This paper explores parallel flow line scheduling environment where similar set of machines with varying capabilities exist. The jobs are required to be processed by every machine available in the assigned line and the processing time of similar machines differs between lines. The complexity increases in parallel flow line scheduling because it combines both the flow shop and parallel machine arrangement. This work concentrates on the considered objective, i.e., minimization of makespan in a large parallel flow line machine setup using three metaheuristics, namely Genetic Algorithm (GA), Simulated Annealing (SA), and Bee Colony Algorithm (ABC). To perform computations, a code with Visual Basic language was developed and run on a personal computer. From the test conducted on a set of randomly generated problems demonstrates that GA outperforms SA for the given objective. Secondly, the solutions of GA are compared with ABC Algorithm and inferred that GA outperforms in all instances. Thirdly, the solutions of SA and ABC algorithm are compared. On comparing the results of various algorithms on the same set of problems considered, it is revealed that GA outperforms other algorithms for the chosen conditions of random samples. To measure the efficiency of algorithms apart from the fitness function, computational time taken to achieve the near optimal sequence in the instances is also considered.