The Research And Implementation Of LDSS Platform Based On Data-Mining

Logistics is defined as a business planning framework for the management of material, service, information and capital flows. It plays an important role in both national economy and social development. Decision-making is most important in logistics. The surviving and developing of a logistic company mainly depends on whether the correct decisions can be made. Thus, logistic companies need a scientific system to help them make decisions. However, most of our country's logistic companies don't have such systems. They usually make their logistical decisions mainly based on their own experience, which is very casual and inefficient.In this paper, the author has designed and developed a Logistic Decision Support System (LDSS), after doing kinds of research on logistics and computer techniques. The system is developed according to the whole logistical industry's requirements, thus it can be used by most of the logistic companies. Three-tier developing framework based on visual studio .net and data mining techniques are adopted, thus the system is very stable and flexible. LDSS has two major functions, one is to maintain daily information of all departments and the other is to support manager's decision-making. Two special decision support sub-system (DSS) is provided. They are distribution center selection DSS and stock Management DSS.The distribution center selection DSS is implemented using genetic algorithm (GA). Considering the special industrial background of logistics, the author has improved the basic GA in three aspects. Firstly, mixing parallel strategy is used in coding variables. Then, adaptive crossing rate which decrease relatively to generations is adopted when crossing individuals. Finally, constraints on the adaptive function are relieved by improved penalty functions. Comparison experiments between this improved GA and the basic GA has shown that improved GA is much more accurate and efficient. Thus, it can solve the problem of best distribution center selection.The stock management DSS is implemented using C4.5 decision tree algorithm in data mining. This algorithm is simple and fast, and it generates rules that are easy to understand. Experiment on collected data sets has proved that C4.5 algorithm can solve this problem commendable.Logistics decision support system is very important in logistic management. It helps to improve the accuracy and efficiency of decision-making, by executing decision-making procedures more scientific. It also helps to improve the company's management ability, thus modern management system is built in the company.