A Data-Driven Storage Assignment Strategy for Automated Pharmacy

Abstract

Automated pharmacy (AP) with random storage location assignment (RSLA) strategy has been widely applied in large hospitals and retail pharmacies. In this paper, an integrated optimization problem of storage location assignment (SLA) and robot arm path planning (RAPP) is considered. For the trade-offs, a Hungarian method (HM)-based storage location assignment (HMSLA) method is proposed for further improving the operation efficiency of AP. Two phases are involved in the proposed method. At phase one, AP is divided into four areas based on BP neural network, and then medicine storage area and specific location are determined through data mining of drug delivery frequency and common combinations. At phase two, HM is applied for optimal scheduling of dispensing multiple medicines. Numerical studies show the proposed method outperforms significantly the traditional RSLA strategy.