Aiming at the assembly sequence planning (ASP) of complex products, the evaluation system for the geometric feasibility, priority constraint, aggregation, redirection, and stability of the assembly sequence was considered. An assembly sequence planning method based on adaptive gravitational search algorithm (AGSA) is proposed. It includes building a mathematical model of algorithm optimization and the coding of assembly products based on the ASP problem. The reverse learning (RL) method is used to initialize the coding product to expand the search range of the initial population, and a new non-repeating exchange (NRE) rule is established to obtain a reasonable optimal value. At the same time, the dynamic adaptive adjustment coefficient (AAC) is used to accelerate the convergence and realize the dynamic adjustment of the assembly sequence until the optimal assembly sequence is obtained. Finally, two examples are given to verify the assembly sequence planning method based on the AGSA, and the superiority of this method is verified by comparing with other algorithms.

针对复杂产品的装配顺序规划（ASP），考虑了装配顺序的几何可行性、优先约束、聚合、重定向和稳定性的评价体系。提出了一种基于自适应引力搜索算法(AGSA)的装配序列规划方法。它包括建立算法优化的数学模型和基于ASP问题的装配产品编码。采用逆向学习（RL）方法初始化编码积，扩大初始种群的搜索范围，建立新的非重复交换（NRE）规则，获得合理的最优值。同时，利用动态自适应调整系数（AAC）加速收敛，实现装配序列的动态调整，直至得到最优装配序列。最后，通过两个实例验证了基于AGSA的装配顺序规划方法，并通过与其他算法的对比验证了该方法的优越性。