Abstract:Network collaborative manufacturing is an important path for the transformation and upgrading of the traditional textile industry. For the core service outsourcing problem, a scheduling optimization method based on an improved NSGA-III algorithm is proposed. The optimization objectives include production cost, total completion time, production quality, customer satisfaction, and resource utilization. A network collaborative manufacturing model is established. On the basis of the NSGA-III algorithm, the Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm is combined to propose a model-solving algorithm with stronger local search capability and convergence ability. The most suitable solution is selected from the Pareto optimal set using a combined weighting method based on Analytic Hierarchy Process (AHP) and Entropy Weight Method (EWM). Through comparative analysis with standard test cases, the results show that the improved NSGA-III algorithm outperforms NSGA-II and NSGA-III in terms of convergence and solution diversity. The effectiveness and superiority of the proposed method are verified through specific network collaborative manufacturing cases.