Collaborative Production Task Allocation Decision for Multi Smart Factory

In this study, we address a multi-factory production task allocation problem in make-to-order manufacturing that simultaneously investigates the two decisions playing critical roles in most firms: order assignment and production time. The model takes into consideration the production time windows, order splitting, and varying capacities of the factories in different periods. To efficiently allocate orders among parallel factories, we improve flexibility and fairness in production task allocation. The research aims to achieve balanced task allocation and cost optimization as primary objectives involving multiple periods and multi-items. We formulate the integrated problem as a mixed-integer nonlinear problem with minimized total costs and assignment fairness among factories. An improved Non-Dominated Sorting Genetic Algorithm (NSGA-II) is developed to address this problem. What's more, we incorporate positive feedback characteristics from the Ant Colony Optimization (ACO) algorithm into the mutation process, enhancing the optimization efficiency of the algorithm while maintaining population diversity. The crossover rate and mutation rate also adaptively improve the convergence, stability, and solution quality of the algorithm. Numerous computational experiments illustrate that the I-NSGA-II algorithm is more capable of finding a feasible solution set in the Pareto space. Finally, simulation results validate the effectiveness of the proposed model and algorithm in achieving satisfactory solutions. Some managerial insights are also explored and reported.