Greedy Algorithm-Based Optimization for Cost-Efficient Supply Chain in Outbound Microchip Logistics

This paper focuses on the optimization of outbound logistics in the microchip supply chain, and compares the performance of the Mixed Integer Linear Programming (MILP) and the greedy heuristic algorithm, based on a real-world dataset of 9,216 orders routed through 15 warehouses, 11 origin ports. The optimization experiments are performed under two strategies, cost-first and wright-first. The results show that the MILP models can reach the global optimal solution that makes it especially suitable for planning scenarios with relatively stable conditions and high accuracy requirements, while the greed algorithm is more acceptable in terms of computational efficiency and scalability. The MILP models are more suited to operational-level logistic planning as it is capable of achieving fast and flexible responses in efficient times, while the greedy algorithms show practical value in real-time operation. The weight-first strategy significantly enhances warehouse utilization and quantity discounts. This research highlights the applicability of the MILP models to strategic long-term scheduling and centralized planning; while the greedy algorithms provide an efficient and feasible alternative for real-time processing and large-scale application environments.