工艺路线的规划决策不仅受制造资源（如机床、刀具、夹具、检具、量具）的影

The planning decision of the process route is not only affected by manufacturing resources (such as machine tools, tools, fixtures, inspection tools, measuring tools), but also restricted by factors such as processing methods, process constraints, and is also related to product types, product batches, and process personnel The design level and even the limitation of the enterprise's process habits make the planning and decision-making of the process route very complicated and become a multi-factor constraint nonlinear programming problem [6,7]. The use of traditional methods (such as Newton's method, gradient descent method, graph theory method) to carry out process route planning and decision-making has the disadvantages of inefficient and time-consuming and easy to fall into local optimal solutions [8-10]; and this static process planning cannot Adapting to the dynamic changes of the enterprise's manufacturing resources has caused the efficiency of the process route to be greatly compromised or even invalidated. In recent years, with the development of artificial intelligence, scholars at home and abroad have proposed a variety of optimization algorithms, such as Hopfield neural network, genetic algorithm, gray correlation method, ant colony optimization algorithm, tabu search, particle swarm algorithm, etc., and achieved a lot of research results [11 -14]; However, due to the defects of each algorithm, and the impact of product diversity and manufacturing resource changes, the use of a single optimization algorithm has certain limitations, and it is difficult to achieve the desired effect. It is necessary to adopt a mixture of algorithms or complementation of different algorithm optimization mechanisms To improve optimization efficiency

The planning decision of the route is not only influenced by the manufacturing resources (e.g. machine tools, tools, fixtures, tools, gypsies), but also by the processing method, process constraints and other factors, but also related to the product type, product batch, as well as the design level of the process personnel and even the enterprise process habits, so that the planning decision of the route becomes very complex and becomes a non-linear planning problem with multi-factor constraints. Using traditional methods (such as Newton method, gradient drop method, graph theory method) to carry out route planning decision-making, there are inefficient and time-consuming, easy to fall into local optimal solution and other defects, and this static process planning can not adapt to the dynamic changes of enterprise manufacturing resources, resulting in the efficiency of the route greatly reduced, or even failed. With the development of artificial intelligence in recent years, scholars at home and abroad have put forward a variety of optimization algorithms, such as Hopfield neural network, genetic algorithm, gray association method, ant colony optimization algorithm, taboo search, particle group algorithm, etc., and obtained a large number of research results. The complementarity of algorithm mixing or different algorithm optimization mechanisms is needed to improve the optimization efficiency

The decision-making of process planning is not only affected by manufacturing resources (such as machine tools, cutting tools, fixtures, inspection tools, measuring tools), but also restricted by processing methods, process constraints and other factors. At the same time, it is also related to product type, product batch size, design level of process personnel and even enterprise process habits Nonlinear programming problems with multi factor constraints [6,7]. Using traditional methods (such as Newton method, gradient descent method, graph theory method) to carry out process planning decision-making has the defects of low efficiency, time-consuming and easy to fall into local optimal solution [8-10]; moreover, this static process planning can not adapt to the dynamic change of enterprise manufacturing resources, resulting in the efficiency of process route greatly reduced or even invalid. In recent years, with the development of artificial intelligence, scholars at home and abroad have proposed a variety of optimization algorithms, such as Hopfield neural network, genetic algorithm, grey correlation method, ant colony optimization algorithm, tabu search, particle swarm optimization algorithm, etc., and a large number of research results have been achieved [11-14]; however, due to the defects of each algorithm, as well as the impact of product diversity and changes in manufacturing resources, a single optimization is adopted The algorithm has some limitations and is difficult to achieve the desired results, so it is necessary to adopt the algorithm mixing or the complementarity of different optimization mechanisms to improve the optimization efficiency<br>