Dynamic distribution optimization method of manufacturing services based on complex network under Industrial Internet platform

With the development of Industrial Internet platforms, the supply of massive manufacturing services leads to more frequent sharing of services among multiple stakeholders. Since Industrial Internet platforms lack optimization methods for the dynamic distribution of multiple manufacturing services, which is key to guaranteeing long-term benefits for manufacturing service stakeholders and ensuring the sustainability of manufacturing systems. To address this problem, based on the digital description of manufacturing services, this paper firstly studies the interaction among manufacturing services to establish manufacturing service complex network. Secondly, a dynamic distribution optimization model of manufacturing services is developed using QoS equal-ratio inverse inference method to predict reasonable distributions among dynamic services under interaction. Then, an improved INSGA-III algorithm is proposed, which includes the dual-population strategy based on crowding distance, the dynamic random opposite learning mechanism and the dual-population adaptive evolution mechanism to solve the problems of inefficient recombination, difficult search, poor diversity and local optima. Finally, two sets of simulation experiments are conducted to simulate the distribution optimization problem of service providers for automotive engine pistons. The improved INSGA-III algorithm's efficacy in addressing the dynamic distribution optimization challenge within manufacturing services is validated through comparative experiments.