考量品項相關性之智慧儲位指派

Abstract

本研究發展智慧儲位指派模型，把品項相關性納入考量。品項相關性是指消費者訂單中品項與品項之間的關聯度，當不同的品項多次同時出現在消費者的訂單中時，品項與品項之間就形成了較高的相關性；當不同的品項幾乎從來沒有同時出現在消費者的訂單中時，品項與品項的相關性就很低。品項的相關性源於消費者的消費形態，消費者少量多樣的消費形態導致其訂單揀貨耗時長，已成為企業提高顧客服務品質的阻礙。影響揀貨作業的因素之一為倉儲系統品項之儲位指派。傳統的儲位指派較多依靠ABC存貨管理法則或是倉儲管理者之主觀經驗，較少將品項之間的關係納入考量。本研究將以總訂單揀貨成本最小化為目標，由關聯法則獲取品項之間的關聯性，再發展智慧儲位指派模型，使其能成為優於EIQ方法的新的儲位指派方法，大幅度縮短總訂單揀貨成本。

This study develops Intelligent Storage Location Assignment (ISLA) algorithm in consideration of correlation between items to minimize order picking costs. The correlation between items is defined by the frequency of simultaneous appearance of items in customers’ orders. When high correlation items were allocated remotely in the warehouse, order picking costs will increase. Hence items with high correlation should be allocated in nearby storage locations to reduce order picking costs. Traditionally, Storage Location Assignment Problem (SLAP) is mainly solved by the subjective experience of warehouse manager or class-based storage policy; little research considers the relationship between items. In the proposed ISLA algorithm, a benefit function is developed to evaluate the fitness of storage location. To evaluate the performance of ISLA, a real-world dataset is used to compare order picking costs under different storage location assignment algorithms. Our results show that the proposed ISLA algorithm regularly outperforms other methods. When used in a multiple-cross-aisle warehouse, the proposed ISLA can reduce picking costs by up to 4.02%.