Optimal Selection of Product Variants to Maximize Market Utility Relative to Assembly Supply Chain Network System Complexity

This study addresses the problem of selecting optimal product variants by maximizing market utility relative to the system complexity of an assembly supply chain (ASC) network. On the supply side, ASC system complexity is quantitatively measured by integrating both ASC network complexity and assembly line complexity using Shannon entropy, which captures the degree of uncertainty and structural intricacy inherent in the system. On the market side, a multinomial logit (MNL) model is employed to estimate customer choice probabilities based on product variants and segment preferences. These probabilities are then aggregated across all market segments to compute the expected market utility of a given set of product variants. To effectively solve the resulting combinatorial optimization problem, a genetic algorithm (GA) is utilized to search for the product variant set that maximizes the utility to complexity ratio. A series of market scenarios, defined by variations in both the mean and variance of market utility distributions, are constructed to evaluate the robustness, adaptability, and performance of the proposed approach. Experimental results reveal that shifts in market utility distribution characteristics significantly affect the optimal selection strategy, underscoring the importance of simultaneously considering both supply side complexity and market side benefits in designing an efficient and competitive product line.