Study On Scheduling Optimization Of Maximizing The Net Present Value Of Projects

In the 1950s United States researchers firstly used network technology, in the 1960s Professor Hua Luogeng introduced this technology to China, and network technology has been rapidly applied to project management and other fields. Network optimization technology is a high-level network planning, and also it is is an important part of research fields in project management.Network optimization contains four different sides, which are fixed duration- resources balance, resources limited-shortest duration, duration-cost exchanged and maximizing the net present value of projects, and the last side in recent years has become a focus in network optimization.Based on the fore-researcher's achievement, some research has been done in this paper on the problem of maximizing the net present value of projects, activity's pattern, limited resources, payment scheduling, contract reward and punishment and other factors which all affect the problem are discussed, and based on the established model, an improved non-linear programming integer model which is suitable for the problem characteristic in this paper is obtained.Particle swarm algorithm is a new optimization approach that developed recently. It has some advantages, that are simpler function, quicker rate of convergence, better global convergence, little requirement to the goal function with better transplantation. According to the discrete feature of variable in the problem discussed in this paper, discrete variable particle swarm algorithm is applied, and it obtains satisfactory results.In order to carry out the algorithm, a main variable which has three parts that are activity-order, payment scheduling and activity pattern is designed. In these three parts, how to solve the activity-order's generation is a problem. Series schedule generation scheme and activity's random priority are applied to generate the activity's order, according to the discrete characteristic of the activity's order, successful experience which is obtained by discrete varable particle swarm algorithm used in Travel Salesman Problem(TSP) is borrowed and the algorithm's position and velocity are redefined, the TSP's generative functions are applied in this paper. During the process of the generation of activity's order, the feasibility of the variable's velocity which is generated by the activity' order is discussed.