Font Size: a A A

Research On The Uncertain Resource-Constrained Project Scheduling Problem

Posted on:2011-11-28Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y YingFull Text:PDF
GTID:1119330332473695Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of manufacturing, information and automation technology, management by projects has been increasingly popular in recent decades Project scheduling has become a core practice of project management both in the field of theoretical research and management practices. According to resource requirements planning (resource constraints), project managers have to determine the start and completion time of each task, and calculate the amount of resources required for each stage, in order to optimize the predetermined objectives. The resource-constrained project scheduling problem (RCPSP) is widely used in industries as well as research fields. Models for the RCPSP are comprehensive, which has been attracting many domestic and foreign scholars.Current RCPSP problem mainly focuses on the deterministic resource constraints, i.e., the resource supply is definite. However, in the actual project practices, for most projects, resource supplies are often uncertain. Uncertainties often lead to variation in the supply of project resources. Current research relies on fixed resources constraints, which often leads to project failure. In recent years, much attention has been paied to the project scheduling problem with uncertain resource constraints.To solve these problems, motivated by some previous studies, this thesis focuses on the RCPSP extended to the environments of uncertain resource constraints, develops mathematical models of the uncertain resource-constrained project scheduling problem (URCPSP); put forward to extend fully resource constraints into account the uncertainty of resources, focusing on constraints for flexible working hours, time-varying resource constraints and uncertain resource constraints RCPSP problem, layers of in-depth analysis, the resource-constrained project scheduling problem (RCPSP) extended to the uncertain environment of resource constraints, achieving the following main contributions: (1) Analyzing resource constraints in the flexible project scheduling problem (flexible RCPSP, FRCPSP), and formulating a complete mathematical model. Flexible resource constraint is referred to the situation where resources upper limits are flexible at the beginning of the project. This thesis presents an mixed ant colony algorithm (MACO_FRCPSP) to quickly find the resources upper limit and their corresponding optimal scheduling of the project duration changes, which can help the project managers make decisions. Computational results show that the proposed project scheduling algorithm is capable to satisfy the practical demand.(2) Analyzing time-varying resource constraints of the RCPSP problem (time-varying RCPSP, TRCPSP) and formulating an exact model. TRCPSP are referred to the RCPSP when arrival time of resources is uncertain, such as raw materials procurement, lease or outsourcing situation. This thesis presents an improved genetic algorithm using the task priority lists as coding schemes (TPLC_TRCPSP),, which can quickly get the resources and the optimal arrival time duration of the project. The computational results verified the efficiency and effectiveness of the proposed algorithm.(3) Analyzing the project scheduling problem in the circumstances of which resources upper limit and arrival time are uncertain, and putting forward a generalized model. This thesis presents a robust optimization model based on genetic algorithms (ROGA_URCPSP), to allow the project to get a kind of uncertainty in the environment, and get the best projects the average duration of the task execution sequence. The computational results verified that the proposed algorithm solves the project scheduling problem with various uncertain resource constraints.(4) Based on the single project scheduling problem, the thesis further studied the uncertainty of resource constraints for multi-project scheduling problem. This thesis proposes an improved genetic algorithm based on random keys encoding of multi-project scheduling program, achieve get the project's multi-project task execution sequence achieving optimal weighted average duration of the project in the environment of uncertain resource constraints. Computational results show that the proposed algorithm can solve the complex multi-project scheduling problem more effectively than traditional methods.
Keywords/Search Tags:project scheduling, multi-project scheduling, uncertain resource constraint, genetic algorithm, ant colony algorithm, robust optimization
PDF Full Text Request
Related items