不同需求特性下多運務員動態分區派遣策略之研究

Dynamic Zoning Strategies for Dispatching Couriers under Different Demand Patterns

本研究考慮在不同時間與空間分布特性的顧客需求型態下，由單一場站指派多位運務員對一固定服務區域內動態產生的顧客進行取件作業的問題。動態需求的時間分布特性包括無尖峰、單尖峰與雙尖峰的情況，而空間分布則考慮均勻與群聚的情況。動態派遣策略的設計包括「動態等待」與「動態分區」兩部分：先完成動態等待的條件後，即進入動態分區以k-medoids法分群，再以Voronoi圖分派每位運務員的責任區域。本研究以系統模擬模式評估不同動態派遣策略之績效，並與傳統之「固定分區」與「不分區」的派遣績效作比較。結果發現在各種測試情境之下，本研究提出結合動態等待與動態分區的派遣策略無論在「營運成本」或「勞役分配」方面，均優於固定分區與不分區的傳統派遣策略。

This research is concerned with the dispatching of multiple couriers in a fixed region with dynamic demand patterns influenced by temporal and spatial characteristics. The temporal characteristics include uniform, single peak-hour and double peak-hour distributions over a day of operation; spatial characteristics include uniform and cluster distributions over the service area. The dynamic dispatching begins with a dynamic wait. Once the waiting condition is met, we start the dynamic zoning process using the k-medoids method to cluster demand points, and then the Voronoi graphs to define the service zone for each courier. Simulation models were built to evaluate the performance of the proposed dynamic dispatching strategy. For comparison the conventional "single zone" and "fixed zone" strategies were also considered. Under various temporal and spatial scenarios tested, the results showed that our proposed dynamic dispatching strategy outperformed conventional strategies in terms of both operation cost and workload distribution.