| 摘要: | 研究指出電子商務的成功因素有許多，其中之一就是完善的物流服務。全球電商龍頭亞馬遜深知此道理，在2014年開始啟用Kiva機器人來改善揀貨作業，以進一步節省人力與快速出貨。Kiva揀貨系統大幅顛覆傳統的揀貨方式，在傳統的揀貨系統，貨架是固定的，揀貨人員必須行至貨架進行揀貨；但是在Kiva揀貨系統，貨架則是可以移動的，Kiva機器人載貨架（亞馬遜稱此貨架為Pod）至揀貨人員面前進行揀貨。自從亞馬遜使用Kiva揀貨系統後，大陸也有幾家公司開始模仿且開發出類似的系統，並成功地被網路電商所使用。如同製造業的「工業4.0智慧自動化製造」，利用Kiva機器人於揀貨自動化將是物流業邁向「物流業4.0智慧自動化物流」的重要里程碑。但是，Kiva揀貨系統的成功不是僅仰賴硬體的創新，同樣重要的是其系統內的作業管理、決策與控制問題的解決，因為這些問題的解決會影響系統的表現。本研究的目的是希望瞭解類Kiva揀貨系統會有哪些作業管理、決策與控制問題；然後研究這些問題，並提出解決方法；最後，建立模擬模式來測試與比較所發展的方法。本研究規劃以三年的時間來進行。第一年計畫已獲補助，且目前進行中，此次申請的是第二年計畫。第一年的研究問題為「訂單選取問題」與「訂單分配（至揀貨工作站）問題」。接續第一年的研究，第二年計畫將針對「Pod分配之揀貨工作站挑選問題」、「Pod分配問題」以及「Pod分配之品項分配問題」來進行探討。「Pod分配之揀貨工作站挑選問題」之目的是如何從可請求Pod的揀貨工作站集合中，找出可優先分配Pod的揀貨工作站。「Pod分配問題」之目的是如何找出一台Pod來分配給可優先分配Pod的揀貨工作站。「Pod分配之品項分配問題」之目的就是如何將Pod上面的品項，分配給揀貨工作站上的未滿足訂單。這三種作業決策問題對揀貨工作站的表現有直接的影響，因此針對這些問題發展優良有效的方法是有其必要的。本研究希望研究成果有實質的應用價值，因此在方法發展上會重視方法於實際環境的有效性與可行性。雖然目前國內廠商尚未有使用或是開發出類Kiva揀貨系統的產品，但是本研究相信類Kiva揀貨系統將會是物流倉庫智慧自動化的未來主流之一。本研究相信本計畫成果將能對國內廠商，未來在使用或是開發類Kiva揀貨系統上有所助益，也因此希望能有機會來順利完成本研究所規劃的三年期計畫。 ;Research shows e-commerce has many success factors and good logistics service is one of them. Amazon, the global e-commerce leader, knows this well. In 2014, it unveiled its eighth-generation logistics center which used Kiva robots to improve order-picking operations. Kiva picking systems drastically change the traditional picking method. In the traditional picking system, shelves are fixed and pickers approach shelves for picking. However, in a Kiva picking system, shelves (also known as pods) are removable. Kiva robots carry shelves to pickers for picking. Since Amazon used the Kiva system, several Chinese companies began to imitate and gradually develop similar systems that have been successfully used by several e-retailers. Similar to the “Industry 4.0’s intelligent and automated manufacturing” in the manufacturing industry, using Kiva robots for picking operations will be an important milestone for the logistics industry toward the “Logistics 4.0’s intelligent and automated logistics center.” The success of Kiva systems is not solely due to the innovation of hardware devices, equally important are the solutions to operations management, decision-making and control problems within the system, as the resolution of these problems greatly affect the system’s performance. The purpose of this project is to study them and propose solutions for them; and conduct simulation experiments to test them. This project plans to study the operations management, decision-making and control problems for a Kiva-like picking system and proposed solutions for them in three years. The first year of the project has been subsidized and is currently in progress. This application is for the second year. The first year's research problems are “order selection” and “order assignment (to picking workstations).” The second year will discuss the “workstation-selection problem of pod assignment”, “pod assignment problem” and “item-allocation problem of pod assignment.” The purpose of the “workstation-selection problem of pod assignment” is to find out which picking workstation has the highest priority to be assigned with a pod. The purpose of the "pod assignment problem" is to find a pod and assign it to the picking workstation that has the highest priority to be assigned with a pod. The purpose of “item-allocation problem of pod assignment” is to allocate items on a pod to unsatisfied orders on the picking workstation. These three problems have a direct impact on the performance of the picking workstation. Therefore, it is necessary to develop excellent and effective methods for them. This project hopes the research results will have substantial application value, thus when developing methods, this project will pay attention to the feasibility and effectiveness of methods in the actual environment. Although, at present, domestic companies have not yet used or developed any Kiva-like picking systems, we believe Kiva-like picking systems will be one of the future mainstreams of intelligent and automated logistics centers. Therefore, we believe the results of this project will benefit domestic companies in their future adoption or development of Kiva-like picking systems and hope to have the opportunity to successfully complete this three-year research project. |
