| Summary: | 碩士 === 中華大學 === 運輸科技與物流管理學系碩士班 === 94 === Recently, the demand of the cold logistics/chain and the multi-temperature distribution has rapidly grown. Carriers must deliver goods with different temperate, such as hot food (over 60℃), normal temperature good (18~25℃), refrigeration food (-2~7℃) and frozen food (under -18℃), to customers. How to distribute multi-temperate goods at the same time and with lower cost becomes an interesting research issue.
This thesis considers two special operational situations: first, carriers utilize the engine-driven frozen truck divided into three parts to hold different temperate goods, and second, carriers utilize the multi-temperature storage box to hold different temperate goods in a general truck. We transfer the previous situations into two Heterogeneous Multi-temperature Fleet Vehicle Routing Problems, HMFVRP1 and HMFVRP2. Then, we also develop a simple heuristic algorithm to solve these HMFVRPs. This heuristic algorithm firstly applies a modified Farthest-start Nearest Neighbor (FNN) method to construct an initial solution, and then improves the initial solution by sequentially executing 2_opt, Or_opt, 1_0, and 1_1 neighborhood searches.
A bank of 168 instances created by modifying the Solomon’s VRPTW benchmark instances, Taillard’s VRP benchmark instances and Homberger’s VRP benchmark instances is used to compare the performance of HMFVRP1 and HMFVRP2. Furthermore, real costs and capacities of different size of trucks are set for these test instances. Experimental results present that, in average, HMFVRP2 performs well than HMFVRP1 in both of fleet size and traveling distance. Such a finding maybe offers an alternative to improve the performance of practical multi-temperate distribution
|
|---|
