Sustainable modular product architecture design by Bi-level leader-follower joint optimization with switching-based meta-heuristic algorithm

Abstract

Abstract Traditional product architecture design is motivated towards structural-functional modularity by a flat map all-in-one (AIO) approach, providing little emphasis on interface complexity and addressing conflicting goals. The resultant modular product architecture (PA) often reduces design robustness and increases complexity for product assembly and recovery. Sometimes, the generated architecture is inconsistent with efficient final assembly, increasing redesign costs. To overcome such problems, this paper proposes a sustainable modular product architecture (SMPA) to reduce assembly complexity and middle-aged discarding of products by enhancing assemblability and recovery. For achieving sustainability in modular design, three sustainable modular drivers—namely, product configuration cost, interface complexity and material similarity—are considered along with several architectural drivers. The key challenge for incorporating sustainability consideration into the existing modularization theories and methods are the heterogeneity between different goals and the decision criteria. To deal with this challenge of decision hierarchy, a bi-level leader-follower joint optimization (LFJO) model is formulated as a Stackelberg game, combining the architectural and sustainability factors of modular design. The main objective is to determine the optimal level of module granularity to reduces assembly cost and enhance product recovery within a coherent framework of total product configuration cost (TPCC) and total interface complexity (TIC). A switching based bi-level meta-heuristic algorithm (SBMA) is developed incorporating Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) to solve the bi-level LFJO problem. A case study of a refrigerator modular design reveals that the joint optimization of product configuration design and interface complexity excels sustainability characteristics of modular design.