Research On Routing Optimization Problem In Fourth Party Logistics With Discounted Costs

With the globalization of supply chain, the transparency of Internet and the personalization of market demand, the third party logistics (3PL) seems to be inadequate, and the fourth party logistics (4PL) appears which is consistent with the times. It is very important for us to research4PL. Because it can help us to guide the development of our business, logistics reform, resources integration and logistics industry.In the study of4PL, routing optimization is a key issue, which includes supply chain management and optimization selection of3PL. Considering the characteristics of the4PL routing problem, a thinking, discounted costs based on multiple tasks, has been proposed. In my opinion, the thinking is an innovation in4PL, our goal is to make the4PL suppliers’ proposal more reasonable.After considering the fourth party logistics, discounted costs and ant colony algorithm, the main work of the paper is as follows:First, a single point to single point routing problem of4PL is discribed by a multi-graph. And a reasonable definition of discounted costs and the mathematical model of4PL routing problem are established.Second, enumeration algorithm, basic ant colony algorithm, ant colony algorithm with dynamic transition probability and pheromone classification improved ant colony algorithm have been used to solve the problem. The effectiveness of the algorithms has been verified by a series of experiments.Finally, a multi-point to multi-point routing problem of4PL has been discussed. Based on the single point to single point routing problem of4PL, two mathematical models are proposed. The analysis shows that the unified model is more suitable which is in line with4PL characteristics. Therefore, the unified mathematical model is used and four kinds of algorithms are used to solve the problem. Through experiments and results analysis, we draw a conclusion:enumeration algorithms can reach the optimal solution, but it’s too complicated for the time, which we cannot accept. Basic ant colony algorithm takes less time, better mean and smaller error, but in large scale, this algorithm easily falls into local optimal solution. Ant colony algorithm with dynamic transition probability can excellently overcomes the stagnation of the basic ant colony algorithm. It gets a better solution. Pheromone classification improved ant colony algorithm is divided into different types by different tasks, it is rational in theory. The experimental results also indicate that this algorithm is better than the basic ant colony algorithm.The research of the paper has a certain reference value to the actual logistics.