Multi-objective double-floor corridor allocation problem with floor loads and separated passages

Current research on multi-floor stereoscopic workshop layouts disregards the effects of floor load constraints and interference with the workers and logistics, which lead to low productivity and escalating transportation risks. This study examines a double-floor corridor allocation problem considering floor loads and separated worker-logistics transportation passages (DFCAP_FLSP). To this end, a mixed-integer programming model involving constraints on facility allocation, multi-type transportation distance computation, and floor loading is constructed to minimize material handling costs, employee walking distances, and floor loading gaps. Subsequently, a multi-objective self-learning memetic algorithm (MOSLMA) is developed to solve the DFCAP_FLSP efficiently. The algorithm employs linear programming to achieve two-stage decoding and utilizes Q-learning to improve local search performance. Comparison experiments conducted with the genetic algorithm, multi-objective particle swarm optimization, and non-dominated sorting genetic algorithm II for the four benchmark instances reveal that MOSLMA achieved the highest percentages of non-inferior solutions close to the Pareto solution (68.75%, 93.18%, 100%, and 87.5%), highlighting its advantages. Finally, the proposed layout and algorithm are applied to the reducer manufacturing workshop, and the scheme comparison indicates the superiority of the proposed layout structure in safety and cost-effectiveness.