Fuzzy Disassembly Planning with Stimuli-Activated Self-Disassembly Workstations

Remanufacturing is vital for sustainability, but its widespread implementation faces obstacles such as labor intensity and quality degradation during disassembly. Stimuli-activated self-disassembly presents an innovative solution, enabling simultaneous component disassembly without damage, thereby reducing labor and energy costs. This cutting-edge approach, when combined with 4D printing, allows products to self-disassemble in response to environmental triggers. In this paper, fuzzy simulation combined with Petri net representation is utilized to evaluate component quality. Additionally, a novel Hebbian learning algorithm is introduced to address real-time disassembly planning challenges. Rapid generation of non-dominated solutions is achieved through a modified learning architecture. The Hebbian learning policy neural network adapts to the fuzziness embedded in triangular membership functions and incorporates the crossover-mutation procedure of the Non-Dominated Sorting Genetic Algorithm (NSGA2). Experimental results demonstrate the efficacy of the proposed algorithm using practical examples, such as Ballpoint pen cases, proving its effectiveness in real-world applications.