A study of the multi-strategy improved dung beetle algorithm and its application to shop floor scheduling problems

Aiming at the problem that the DBO algorithm is slow to converge at the initial stage and prone to fall into local optimality at the later stage, the multi-strategy improved dung beetle algorithm (MDBO) is proposed to enhance the global search and local exploration ability through the global optimal solution guidance, the adaptive mechanism, and the Cauchy-Gaussian variant. It is applied to the Distributed Flexible Job Shop Scheduling Problem (DFJSPT) considering transportation time with two-dimensional coding and plug-in decoding, and combined with mixed population initialization, discrete-continuous transformation and tournament selection strategies to optimize the solution process. Experiments show that distributed scheduling has significant advantages and transportation time has a large impact on production scheduling. In 20 extended algorithms, the improved IMODBO algorithm is compared with NSGA - II and MOPSO, highlighting its superiority in multi-objective solving. For example, the IMODBO algorithm improves the production efficiency by 11.19%, reduces the total energy consumption by 7.50%, reduces the bottleneck machine load by 16.92%, and improves the machine utilization rate by 4.78%, which verifies its excellent performance and promising prospect in the practical application of the IMODBO algorithm.