Biogeography-based optimization algorithm for large-scale multistage batch plant scheduling

Abstract

The batch process is characterized by many varieties, small batches, redundant production equipment, flexible production process, and high-added-value products. This process is widely used in chemical, plastic, rubber, pharmaceutical, fine chemical, metallurgical, steel, food, and other industries. The optimized scheduling scheme of the batch process can effectively enhance enterprise competitiveness and improve economic benefits. Multistage Multiproduct Scheduling Problem (MMSP) is an important branch of batch scheduling problems. It is difficult to solve MMSP within a reasonable time by traditional mathematical programming, because once the scale of scheduling problems increases, the solution space expands exponentially. This study proposes a metaheuristic approach based on a time key biogeography-based optimization algorithm to solve MMSP. This new time key representation contains two vectors, which represent the processing sequence and equipment allocation of orders respectively. In accordance with the time information in the new representation, we add the preference of equipment processing time to migration and calculate the probability of every mutation value. In addition, the elite solution is combined with the active scheduling technique and modified Nawaz-Enscore-Ham (NEH) algorithm to improve the search accuracy of the proposed algorithm. To test the performance of Improved Time Key Biogeography-Based Optimization (Improved-TKBBO) algorithm, its results are compared with computational results of mathematical programming, Genetic Algorithm (GA), and Line-up Competition Algorithm (LCA). Simulation results show that the proposed Improved-TKBBO can solve the large-scale MMSP with non-identical parallel units effectively.