<![CDATA[Micro, small, and medium enterprises (MSMEs) are important economic drivers for Indonesia, especially in labor-intensive sectors like footwear manufacturing. MSMEs, though, face acute logistical problems because of heterogeneous customer demand, limited production capacity, and ever-increasing transportation costs. Few existing works have focused on monthly logistics planning for MSMEs in developing countries with realistic costing and demand structures. To develop and analyze a Genetic Algorithm (GA) optimization model to maximize profit within a constrained monthly footwear profit distribution network. To achieve this, we needed to assess how multi-retailer product allocation balance could be achieved with minimum operational constraints such as production caps, cost-efficient logistics, and streamlined processes. This study employed a quantitative experimental design approach and implemented a GA with real-valued chromosome representation, tournament selection, single-point crossover, and Gaussian mutation. The model was built using real data from a footwear MSME operating in the Lamongan and Tulungagung regions of Indonesia. The algorithm was implemented using Python and tested for reliability with 10 executed validations for independence. Within 60 generations, the GA maintained consistent convergence and achieved a final fitness value with a coefficient of variation of 0.24%. The optimized allocation achieved a net profit margin of 15.22% while utilizing the available production capacity (600 units/month). Because of increased profit contribution, greater-distance wholesale customers were served first despite incurring higher transport costs. The model had no constraint violation and reduced transportation costs to 1.45% of total revenue. Using GA to address multi-objective distribution challenges in the context of MSMEs appeared to have positive results, confirming the effectiveness of this approach. The proposed approach helps frame and guide critical allocation and routing decisions, which can be made within the boundaries of operational constraints. Further work is needed to incorporate stochastic demand modelling and multi-objective problem extensions and seek real-time application to bolster support for decision-making in dynamic scenarios.]]>
Copyrights © 2025
