隨著電子商務快速發展與倉儲自動化技術持續成熟，物流中心面臨日益嚴峻的揀貨效率挑戰。在以LocusBots為代表的人機協作揀貨系統中，機器人需依據倉儲狀態動態選取作業區域與區塊，其決策品質對整體系統績效具有關鍵影響，然而目前針對同時考量距離與資源負載之雙因子派工法則，尚缺乏完整之實證探討。 本研究以類LocusBots人機協作揀貨系統為研究對象，除既有之單屬性派工法則外，另提出結合距離與已配給Robot數量之雙因子法則，於Picker與Robot數量比例為1:1至1:5之五種資源配置下進行模擬實驗，透過多因子變異數分析（ANOVA）與Duncan多重比較法，就系統總執行時間、平均訂單延遲時間、平均單趟旅行時間及平均批次訂單週期時間四項績效指標進行系統性分析。結果顯示，Zone選取法則為影響系統績效最關鍵之決策層次，本研究提出之雙因子法則MAXZDRN於四項指標上均表現最佳，較單純以距離或單純以資源負載為依據之法則更具全面優勢；Block層級則以距離導向法則與雙因子法則MAXBDRN表現最佳，三者效能相近。此外，Robot數量增加存在邊際效益遞減現象，且Robot個體旅途效率與系統整體效率之間存在取捨關係。 本研究成果可為倉儲業者在派工法則設計與資源配置決策上提供參考依據，並為後續人機協作揀貨優化研究奠定基礎。;With the rapid growth of e-commerce and the continued maturation of warehouse automation technologies， distribution centers face increasingly stringent challenges in order-picking efficiency. In human-robot collaborative picking systems represented by LocusBots， robots must dynamically select operating zones and task blocks based on warehouse conditions， and the quality of these decisions has a critical impact on overall system performance. However， systematic comparative research on dual-attribute dispatching rules that jointly consider distance and resource load remains limited. This study uses a LocusBots-style human-robot collaborative picking system as the research subject. In addition to existing single-attribute dispatching rules， this study proposes dual-attribute rules that combine distance with the number of robots already assigned. Simulation experiments were conducted under five Picker-to-Robot resource configurations ranging from 1:1 to 1:5， and multi-factor analysis of variance (ANOVA) together with Duncan′s multiple range test were applied to systematically analyze four performance indicators: Total System Time， Average Order Tardiness， Average Trip Time， and Average Flow Time. The results show that the Zone selection rule is the most critical decision level affecting system performance; the proposed dual-attribute rule， MAXZDRN， achieved the best performance across all four indicators， demonstrating a more comprehensive advantage than rules based solely on distance or solely on resource load. At the Block level， distance-oriented rules and the dual-attribute rule MAXBDRN performed best， with the three rules showing comparable effectiveness. In addition， increasing the number of robots exhibits diminishing marginal returns， and a trade-off exists between individual robot trip efficiency and overall system efficiency. The findings of this study provide practical reference for warehouse operators in dispatching rule design and resource configuration decisions， and lay a foundation for future research on human-robot collaborative picking optimization.
