隨著互聯網和科學技術的進步，供應鏈節點分工的深入，傳統的物流服務業正在發生着深刻的顛覆性變化，專業化的第三方物流公司也逐步往平臺化、智能化的方向快速發展。鋼鐵出廠物流是典型的傳統物流服務業，越來越多的鋼廠將出廠物流外包給專業化的第三方物流企業，在這些企業中，不少正在凸顯的問題成為了企業發展過程中難以逾越的障礙，調度公平性問題就是其中一個。
企業物流調度實踐中的調度不公平現象，為企業的管理帶來了困難，駕駛員離職比較頻繁。本研究經過深入訪談和調研，梳理了鋼鐵出廠物流業務的特點和流程，在此基礎上提出了調度公平性問題。找到問題的根源“好活”“差活”，以及如何對它們進行定量化的研究是解決調度公平性問題的關鍵。本研究在對關於運輸任務的大量調研問卷的分析基礎上，發現駕駛員最關注運輸任務單位時間內為自己帶來了多少利潤，因而提出了運輸任務效益值、駕駛員效益和整體效益值的概念，以此定量化描述運輸任務和運輸任務在駕駛員中的分派情況。隨後將常見的公平性評價方法如賈恩公平性指數、基於信息熵的公平性評價、基尼係數等評價方法引入到調度公平性的評價與衡量工作中，提出了衡量物流調度公平性水平的調度公平性指數。
科學組織待發貨物以及合理安排運輸車輛是物流調度的使命，通過合理的調度流程與調度方法來保證調度工作對駕駛員相對公正公平是研究調度公平性問題的最終目的。本研究借鑒了公平理論的程序（過程）公平、結果公平的理念，結合H企業的鋼鐵出廠物流調度實踐，提出了面向公平性的兩階段調度法，把一個完整的調度工作分為配載計畫階段和任務分派階段，以此來緩解調度公平性問題帶來的衝突。配載計畫階段主要是將待出廠的任務提前做好規劃，生成一車車可分派給駕駛員的運輸任務，任務分派則主要是對配載計畫階段生成的運輸任務進行合理地分派。接著結合H企業的實證研究，對兩個階段分別進行了改進：在配載計畫階段，建立了基於均衡性的配載計畫數學模型，並設計了遺傳算法進行求解；在任務分派階段，提出了最佳公平性匹配的任務分派方法，建立了對應的數學模型，並對最佳公平性匹配的任務分派流程和步驟進行了詳細的闡述。
本研究最後，對配載計畫、任務分派兩個階段的不同種方法組合成的四種具體的兩階段調度法進行了模擬，運用了調度公平性評價方法對模擬結果中的駕駛員效益值的分佈進行了調度公平性指數的計算。評價結果證明本研究提出的兩階段調度法能夠有效提升調度過程和結果的公平性，對兩階段調度法的改進也是有效的改進。

Outbound Logistics in Steel Mill is the typical traditional logistics industry. Nowadays, more and more steel manufacturers outsource their outbound logistics services to the professional third party logistics enterprises. With the progress of Internet and technology, all the node companies of the supply chain involve deeper and deeper cooperation, leading to a huge change in the logistics industry: professional and specialized third party logistics enterprises are making large advancement towards “platform” and “intellectualization”. “Platform” means that those logistics enterprises need to make full use of social logistics resources, especially to encourage a large number of personal trucks to join the platform, directly providing high quality logistics service to the customers. In order to realize the sustainable operation of platform, logistics enterprises are faced with lots of problems, of which the most significant one is to achieve the fairness of logistics flatten scheduling process.
During the survey of the fairness of logistics scheduling process, the author found out that there exist many unfair cases in outbound logistics process. Those unfair cases not only caused much difficulty to company management, but also resulted in the frequent resignations of the truck drivers who felt being treated unfairly. After numbers of in-depth interviews and research, the author summarized the characteristics and the whole process of steel outbound logistics, then the author described the issue of unfairness and focused on the significance of the unfairness existed in steel logistics scheduling process. By digging into the key factors that caused the unfairness, the author found out that whether the transport task is “good” or “bad” causes the fairness of the work. Thus, how to measure the fairness of every transport task is the first step to solve the issue of unfairness. In the survey with a large number of questionnaires, the author came to realize that the truck drivers focus on the different income brought by different transport tasks, with the same working hours. So the author came up with three variables: the efficiency of the transport task the efficiency of the truck driver, and the efficiency of whole transport tasks, to measure every transport task and show their distribution among truck drivers. In addition, the author used three common evaluation indicators, Jain’s Fairness Index, Entropy-based Fairness Evaluation and Gini Coefficient to build the logistics scheduling fairness evaluation model.
Scientifically and reasonably arranging the goods to be delivered and dispatching the trucks for transportation are the two important parts of logistics scheduling. Therefore, we need to use scientific and reasonable methods to assign every transport task to reduce the unfairness of logistics scheduling. This paper refers to the procedural fairness and outcome fairness of the fairness theory, combining the practice of the logistics scheduling process of H Company. As a result, the author came up with a two-stage logistics scheduling method based on fairness. To be specific, we need to divide the whole logistics scheduling work into two stages, stowage planning stage and tasks assigning stage. The first stage is the stowage planning stage, during which all the transport tasks need to be planned well for the next stage. The second stage is tasks assigning stage, during which all the transport tasks need to be fairly assigned among truck drivers. The author did the empirical analyses for both of the two stages, using the case of H Company. For the stowage planning stage, the model with balance-based constraints and the improved genetic algorithm is proposed. For the tasks assigning stage, the author put forward the fairest method to assign every transport task based on the mathematical model and specifically explained the procedures of the task assigning process.
At the last part of this paper, the author simulated the four different conditions using the two-stage logistics scheduling method, based on two different stowage planning methods and two different tasks assigning methods. What’s more, the author calculated the fairness indicator of logistics scheduling process by the efficiency of truck drivers and the efficiency of all transport tasks. The results showed that the two-step logistics scheduling method raised in this paper can improve quite a lot the fairness of the logistics scheduling process. The outcome in this paper has been proved and well applied in the case of H Company, not only preparing the basis of the platform operation and the scientific management methods to H Company, but also giving some positive insight to modern logistics enterprises.