物流中心中，揀貨由於重複性以及勞動密集性耗費大量人力因此備受重視。但目前研究大部分皆只有解決訂單分批(Batching)或是最佳路徑規劃(Routing methods)。學者提出一二階段之啟發式演算法，以解決兩種問題(joint order batching and sequencing problem)結合之排程架構，並透過小樣本之實驗進行比較，雖然計算較為複雜，但確實可得到相當大之改善。本研究考量物流中心揀貨作業的相關特性，利用整合粒子群最佳化演算法和蟻群演算法規劃訂單分批與揀貨順序的配置，使規劃後之揀貨距離(total travel distance)最小化。首先，本研究將探討過去在揀貨最佳化的研究，以作為此演算法的依據，進而提出揀貨時間最小化的整合式演算法，透過詳細說明演算法之流程與邏輯。最後，本研究所提出之演算法將與線性規劃求解的方式進行比較與效益分析。 Order picking is the most costly operation in the warehouse because it is labor-intensive and repetitive. However, most common order picking research focus on order batching or picker routing and both sub-problem are NP-hard. Therefore, this research purposes a hybrid algorithm for solving the joint batch picking and picker routing problem considering batch size, order allocation in a batch, and traveling distance. The core of the hybrid algorithm consists of the particle swarm optimization (PSO) algorithm and ant colony optimization (ACO) algorithms. The PSO algorithm finds the best batch picking plan by minimizing the sum of the traveling distance. The ACO searches for the most effective traveling path for each batch. The result shows that the hybrid algorithm is the more efficient methods than the optimal solution and the current industry practice in terms of solution quality and computational efficiency.
