本研究主要討論現代化物流的演進以及以人工智能輔助運輸路徑規劃。目前在醫藥物流產業由於運送之醫藥產品對溫度及保存條件較為嚴格，配送點有諸多特殊要求等等限制，故物流中心之焦點在前幾年仍著重在優良運輸規範 (GDP，Good distribution practice) 之認證而有關運輸排程的自動化發展較慢，故大部分仍採用傳統人工分派路線方式為主。本文透過對醫藥及物流產業的解析、GDP、溫控設備及自動化倉儲的說明目前醫藥物流的現代化歷程，以及透過個案公司營運流程解釋從訂單到出貨相關作業進而帶出運輸規劃的挑戰。 個案公司與市場人工智能平台供應商合作開發平台，進行實車驗證及最佳化模擬驗證，以期透過系統導入後能達到包含減少支援車輛之使用、符合勞基法工時及降低營運成本與碳排放。經過驗證發現此系統對於以上三項目標均有產生效益，惟相關變數包含作業時間、停等時間、出發時間需要配送司機予以反饋以提升系統預估精準度。此系統除了對於成本上有效益外，對於管理上也能幫助個案公司降低主管與配送員的資訊落差以及改善資深與資淺的司機配送效能落差。;The research is to discuss the progress on modern logistic service and how AI powered routing optimization improve the transportation and delivery efficiency。The pharma and healthcare industry is highly governed by health authority for patient safety， so the requirements related to temperature management、monitoring and storage condition are still strict to ensure the product quality and force the service provide to focus on Good Distribution Practice with less effort on system automation， especially automated routing system that probably is one of the key reason now most of pharma distributor remain traditional manual routing approach。 We hope to share how the pharma logistics has been improved via the perspective of pharma and logistic industry、GDP regulations 、temperature controlled facility and automated warehouse， and to point out the challenge of transportation planning through the operation process from order processing to dispatch per Zuellig Pharma case. The company in this case is working with a system supplier for AI powered routing optimization to develop a platform to conduct practical testing with fleet on one district in north area and validation of optimal stimulation for all deliveries in north Taiwan， to check if the benefits including reduction on quantity of supportive trucks 、conforming working hour required by labor act and decreasing the operational cost and carbon emission。After verification， we observed the positive results on above 3 objectives， but the variable factors including operation time、waiting time、departure time ect. would be better to be updated per deliveryman advice to improve the accuracy on estimation。The system is not only to reduce delivery cost but also to minimize the information gap between deliveryman and supervisor form management point of view to improve the performance gap between senior deliveryman and junior deliveryman.
