The Study On Critical Chain Project Scheduling Method Based Heuristic Algorithm

The Critical Chain Method (CCM) that is based on Theory of Constraints (TOC) isan approach of an approach to project scheduling plan, which are the important researchcontents in the field of project management in recent years. Compared with traditionalproject management, it takes the project as a whole in the view of system, target to theoverall optimal of the project. Based on the method of Critical Chain, in this paper weconsider the problem with overall optimal of the project. To solve the problem, weestablish the models and design heuristic algorithm and intelligent optimizationalgorithms to the different models. Computational results show our algorithms giveeffectiveness.In this research, we firstly summarize the critical chain approach applied to RCPSPand describe the principle and method. Then, we propose single-mode resourceconstrained critical chain project scheduling problem model and multi-mode resourceconstrained critical chain project scheduling problem model, both with goal of overalloptimal of the project.For single-mode resource-constrained critical chain project scheduling problemmodel, we design a heuristic algorithms based on priority rules and compare theperformance of different priority rules in the critical chain scheduling. In order to getbetter solution, we design a new genetic algorithm. In the paper, we describe principle,decoding, scheme, fitness and genetic operators in detail.For single-mode resource-constrained critical chain project scheduling problem model,considering the non-renewable resource, we present a two-stage solution, whichconverts multi-mode critical chain project scheduling problem into single-mode criticalchain project scheduling problem. Under the guidance of such thought, we designed aheuristic algorithms based on the model-priority rules and a genetic algorithms.We compare the different parameter setting and obtain the best configuration in the GA. In addition, we test and verify all of the algorithms in this paper by the projectexamples from the PSPLIB.