To address the issues of complex physical modeling, poor adaptability, and insufficient algorithm generalization in the current selective disassembly sequence planning problem, a structured heterogeneous graph modeling method was proposed, which combined with an adaptive proximal policy optimization algorithm to achieve efficient disassembly sequence optimization. Through the structured heterogeneous graph modeling, the multi-constraint relationships of the product components were unified, providing a more expressive state representation for subsequent optimization. Additionally, in the optimization algorithm, advantage function normalization and entropy regularization mechanism were introduced to standardize the data distribution inconsistency caused by dimensional differences across different training stages, while adaptively adjusting the exploration intensity during the training processes to enhance the model's training stability and generalization ability. Experimental results show that the introduction to advantage function normalization significantly improves the algorithm's convergence speed and training stability, while the entropy regularization mechanism enhances the algorithm's exploration ability. Compared with traditional deep reinforcement learning algorithms, the proposed method performes better in terms of convergence and the quality of the optimal policy.
SDSP问题是指在产品存在复杂约束的条件下,根据具体情况选择性拆卸部分零部件,以实现最大化拆卸利润或最小化拆卸时间,并规划出最优或近似最优的可行拆卸序列。其中,首要任务是对产品复杂结构与约束关系进行有效建模,从而为后续策略优化提供准确的结构信息支持。现有研究多采用与或图或Petri网等方式,将产品结构转化为约束矩阵,以表征零部件之间的物理连接和逻辑依赖关系。然而,与或图仅能表示单一约束关系且节点本身缺乏承载零部件属性的能力,需要借助额外的矩阵存储利润、成本和拆卸时间等信息;Petri网则在零部件数量增加时,其库所与变迁数量会呈现指数级增长,难以适应大规模拆卸建模需求。针对上述问题,本文引入异构图(heterogeneous graph,HG)作为拆卸信息的建模方法,通过不同类型的边刻画零部件之间的多种约束关系,并结合零部件属性信息构建结构化的产品拆卸异构图(disassembly heterogeneous graph,DHG)。图1所示为电池案例的DHG。该图由三元组G={ V,Ex,Ey }表示,其中,G表示异构图, V 为节点集合, Ex 为物理接触约束边, Ey 为空间依赖约束。
2)动作空间A由DHG中的节点及约束关系定义,每个动作对应于选择一个零部件节点执行拆卸操作。并非所有节点在任意时刻都可被选择,其可行性取决于当前拆卸状态以及所施加的约束关系。具体而言,物理接触约束 Ex 要求其前驱零部件全部拆除后方可执行拆卸;空间依赖约束 Ey 则只需部分前驱零部件拆除即可。基于此,通过判断DHG中节点的约束满足情况,可在每个状态下确定当前有效的动作集合。这种动态机制能够在不同拆卸状态下实时排除实际不可执行的动作,从而在动作空间维度不断变化的条件下保持策略网络的泛化能力。
ZHUZhuoyue, XUZhigang, SHENWeidong, et al. Selective-disassembly Sequence Planning Based on Genetic-bat Algorithm[J].Journal of Zhejiang University(Engineering Science), 2018, 52(11):2120-2127.
[4]
GUOH, ZHANGL, RENY, et al. Optimizing a Stochastic Disassembly Line Balancing Problem with Task Failure via a Hybrid Variable Neighborhood Descent-artificial Bee Colony Algorithm[J]. International Journal of Production Research, 2023, 61(7): 2307-2321.
GUOHongfei, FU Wenjie LI Leixiao, et al. Research Progress on Optimization-based Disassembly Sequence Planning[J]. Computer Engineering and Applications, 2025,61(11):51-66.
RENYaping, GUOHongfei, ZHANGChaoyong, et al. Energy Consumption Optimization of Selective Disassembly Planning Considering Product Embodied Energy during Manufacturing[J]. Journal of Mechanical Engineering, 2021, 57(6): 200-210.
[11]
RENY, GUOH, LIY, et al. A Self-adaptive Learning Approach for Uncertain Disassembly Planning Based on Extended Petri Net[J]. IEEE Transactions on Industrial Informatics, 2023, 19(12): 11889-11897.
[12]
HUY, LIUC, ZHANGM, et al. An Ontology and Rule-based Method for Human-robot Collaborative Disassembly Planning in Smart Remanufacturing[J]. Robotics and Computer-Integrated Manufacturing, 2024, 89: 102766.
WANGLei, HEChen, CAOJianhua, et al. Failure Status Based Planning Method for Selective Disassembly Sequence of Used Machinery Products[J]. Modern Manufacturing Engineering, 2022(5): 145-152.
TIANYongting, ZHANGXiufen, XUJinfang, et al. Selective Parallel Disassembly Sequence Planning Method for Remanufacturing[J]. Journal of Computer-Aided Design & Computer Graphics, 2018, 30(3): 531-539.
[17]
ZHAOX, LIC, TANGY, et al. Reinforcement Learning-based Selective Disassembly Sequence Planning for the End-of-life Products with Structure Uncertainty[J]. IEEE Robotics and Automation Letters, 2021, 6(4): 7807-7814.
[18]
XIAOJ, ZHANGZ, TERZIS, et al. Dynamic Task Allocations with Q-learning Based Particle Swarm Optimization for Human-robot Collaboration Disassembly of Electric Vehicle Battery Recycling[J]. Computers & Industrial Engineering, 2025, 204: 111133.
[19]
LIY, ZHUW, GUOJ, et al. Multi-objective Collaborative Optimization of Green Disassembly Planning and Recovery Option Decision Considering the Learning Effect[J]. Journal of Manufacturing Systems, 2025, 80: 324-343.
[20]
PENGY, LIW, ZHOUY, et al. Dynamic Disassembly Planning of End-of-life Products for Human-robot Collaboration Enabled by Multi-agent Deep Reinforcement Learning[J]. IEEE Transactions on Automation Science and Engineering, 2025, 22: 13907-13919.
[21]
RENY, MENGL, TIANG, et al. An Efficient M-step Lookahead Rollout Algorithm for Profit-oriented Selective Disassembly Sequence Planning with Operation Stochastic Failure[J]. Engineering Applications of Artificial Intelligence, 2025, 145: 110173.
[22]
CHUX, LIL, ZHAOF, et al. Disassembly Time Estimation for Used Smartphones Based on Maynard Operation Sequence Technology[J]. Computers & Industrial Engineering, 2024, 193: 110291.
LIUYifei, LIXiaoshuai, YANGJun’an, et al. SANER-PPO Algorithm-based Jamming Resource Allocation for UAV Swarm[J]. Control and Decision, 2024, 39(12): 3937-3945.
[25]
ALFARO-ALGABAM, RAMIREZF J. Techno-economic and Environmental Disassembly Planning of Lithium-ion Electric Vehicle Battery Packs for Remanufacturing[J]. Resources, Conservation and Recycling, 2020, 154: 104461.
[26]
WANGK, LIX, GAOL, et al. A Discrete Artificial Bee Colony Algorithm for Multi Objective Disassembly Line Balancing of End-of-life Products[J]. IEEE Transactions on Cybernetics, 2022, 52(8): 7415-7426.