Multi-Objective Optimization Method Based On Variable Weights For Flow Allocation In Logistics Network

The efficiency and effectiveness of logistics network are directly dependent on the network topology and the allocation method of flow in logistics nodes and transportation lines of the network. Once the logistics network topology is formed, it will not be changed in quite a long period of time. However, the flow allocation method for logistics network will be influenced by supply, demand, network capacity and many other factors. In practical application, every node or line of the logistics is related to many evaluation criterias. When allocating freight flow in logsitics network, cost is not the only consideration, other fators such as transport time, distance and reliability should be also considered. Therefore, in order to meet the need of freights’transportation, storage and delivery from suppliers to the customers and obtain the maximum efficiency and benefit of the logistics network, how to allocate the flow in the logistics networks with fixed topology scientifically and rationally is an important issue worthy of depth study.On the basis of the review of existing researches, the problems of flow allocation in a tow level logistics network with many to many structure are studied in this dissertation, and the main contents of the dissertation are as follows:1. Variable weight method for logistics network is proposed. Aiming at the diversity of optimization objective, the limitations and uncertainty of the decision makers’preferences about the problem of flow allocation in logistics network, a logistics network variable weight method is proposed, including representation for the factors of logistics network, such as logistics nodes, transportation lines, sub-paths and paths, and calculation method for their variable weight state values. First of all, values of optimization objective functions are evaluated by membership function; then, using state variable weight function which has considered the decision makers’preferences and variable weight sum method, the state values of logistics nodes, transportation lines, sub-paths and paths are obtained by integrating all of objective evaluations in stages, and operation of choosing the best path is in accordance with the state values of all paths; finally, the method for flow allocation which based on the logistics network variable weight method is proposed. 2. Multi-objective optimization problems for flow allocation in logistics network of which intermediate nodes has no supply capability are studied. For a two level logistics network with fixed many to many topology, aiming at the two situations that capacity of transportation lines are certain and stochastic, the problem of flow allocation is studied based on the constraints of transportation lines’and logistics nodes’capacity with the minimum logistics cost which is defined as a concave function because of the consideration of economics of scale, the minimum maximal delivery time for single path and the maximum logistics network’s utilization ratio when capacity is certain or the maximum logistics network’s reliability when capacity is stochastic as optimal objectives. Firstly, the mathematical models of the problem in both cases are built; and then, due to multi-objective optimization problem for flow allocation in logistics network is belong to NP-hard problem, the number of network paths will increase exponentially with the network scale growing. So, a Genetic Algorithm based on logistics network variable weight model is proposed to solve the problem; finally, a numerical example is given to verify the effectiveness of the algorithm.3. Multi-objective optimization problems for flow allocation in logistics network of which intermediate nodes has supply capability are studied. Based on the work of section2, the supply capability of intermediate nodes is considered. Two situations are taken into account that capacity of both transportation lines and logistics nodes are certain and transport lines, stochastic. Aiming at the two situations that capacity of transportation lines and logistics nodes are certain and that of transportation lines, intermediate nodes and supply nodes are stochastic, the problem of flow allocation is studied based on the constraints of transportation lines’and logistics nodes’capacity with the minimum logistics cost, the minimum maximal delivery time for single path and the maximum logistics network’s utilization ratio when capacity is certain or the maximum logistics network’s reliability when capacity is stochastic as optimal objectives. Firstly, the mathematical models of the problem in both cases are built; then, the solution is proposed that logistics network is decomposed into lower and upper level network which are again decomposed into some many-to-one sub-network, and the problem of flow allocation for each sub-network can be solved by the method proposed in section1, and then the problems of lower and upper level network can be solved successively by the dynamic programming method; finally, a numerical example is given to verify the effectiveness of the solution.