A binary coding genetic algorithm for multi-purpose process scheduling: A case study

This paper presents a novel genetic algorithm (GA) for the scheduling of a typical multi-purpose batch plant with a network structure. Multi-purpose process scheduling is more difficult to deal with compared to single-stage or multi-stage process scheduling. A large amount of literature on this problem has been published and nearly all of the authors used mathematical programming (MP) methods for solution. In the MP methods, a huge number of binary variables, as well as numerous constraints to consider mass balance and sequencing of batches in space/time dimensions, are needed for the large-size problem, which leads to very long computational time. In the proposed GA, only a small part of the binary variables are selected to code into binary chromosomes, which is realized through the identification of crucial products/tasks/units. Due to the logical heuristics utilized to decode a chromosome into a schedule, only the feasible solution space is searched. Our genetic algorithm has first been devised with particular crossover for makespan minimization and then adjusted for production maximization.