Abstract

In the realm of hybrid flow shop environments, existing literature predominantly overlooks a practical scenario that jobs with different constraints often need to be simultaneously processed. This article addresses a constrained hybrid flow shop scheduling problem (CHFSP) in which certain jobs are bound by a common deadline while other jobs are not. We first propose constructive heuristics to generate initial solutions, completely based on the characteristics of the problem. Additionally, we present a variable-representation discrete artificial bee colony algorithm (VRDABC). The algorithm introduces a novel variable solution representation along with a tailored decoding method, allowing an extensive exploration of the solution space while maintaining a remarkable search speed. By leveraging the distinctive solution representation, we design an insertion operation, a multi-phase evolutionary process, and an efficient phase-transition mechanism, thereby enabling the VRDABC to successfully address this problem. Experimental validations, including algorithmic components analysis and algorithmic comparison, confirm the effectiveness and potential of our approach in producing high-quality solutions. Through the theoretical explanation and experimental analysis, the VRDABC presents an influential contribution to the constrained hybrid flow shop scheduling problem.