Research On The Optimization Of Resources Constrained Prolect Scheduling Problem Based On Genetic Algorithm

Nowadays, with the development of the economic globalization, market competition in Civil Engineering is more and more fierce,The traditional steady-type environment for the construction has evolved into dynamic-type. The enterprise has encounter with the fast, continual changed environment. Thus the request for project management becomes higher than before. In the case of resource-constrained, it is a one of an important challenge for the construction enterprise to arrange the schedule suitably. Resource-constrained project scheduling problem (RCPSP) is a typical issue in project schedule. It's a NP hard problem. It is research on the activities of the schedule that the logical relationship and the resource constraints are satisfied while optimizing the managerial objective and makes the time limit shortest. Resource-constrained project scheduling problem has become a hot topic management of project schedule. Therefore, we use the Genetic Algorithm to study the RCPSP on the basis of dynamic characteristic of project that non-fixed total duration, and introduced the solution process of RCPSP in detail.Based on the analyzing of the development of the optimization algorithm at home and abroad and some basic theory. Then on the basis of defining the characteristic of project, according to the principle of project scheduling in resource restriction, the mathematics model of the resource-constrained project scheduling is established. The objective function is to minimal the total time limited. Then we discussed the basic principle and the main steps of the Genetic Algorithm. Base on this, a Genetic Algorithm that aimed at improving operator was introduced, we designed a monomial Genetic Algorithm for the model of RCPSP. And wrote programs based on MATLAB. Finally, in order to confirm the validity of the model and the algorithm, we took an experiment based on Patterson110 instances. The experimental result indicates that the model and the algorithm proposed in this dissertation are feasible. The dissertation also compares with the result of project software witch based on the heuristic algorithm and finds out that the results of the GA are better. So the effectiveness of this algorithm is confirmed.