The distributed permutation flowshop scheduling problem: A genetic algorithm approach

We consider solving a distributed permutation flowshop scheduling problem (DPFSP) with the objective of minimizing makespan. The problem has two dimensions: assigning jobs to factories and scheduling the jobs assigned to each factory. We use GA to solve this problem. In the proposed algorithm we employ some standard techniques like one point crossover and swap mutation. Computational experiments show that the proposed GA produces improved results than original ones but not powerful enough to produce better ones than the known best solutions. Based on this study we will redesign the standard GA implementation by using structural information from the problem. Combining GA with constraint programming and other heuristics to design hybrid search algorithms will also be a valuable direction for further study.