Multi-objective Optimization For Resource-Constrained Large-scale Projuect Scheduling

In the recent decades, researchers and practitioners witness a boom in large-scale projects which are characterized by high uncertainty and increasing complexity. How to formulate and arrange a reasonable schedule plan for these large complex projects has long been an issue in the domin of project management. With more constraints and multiple objectives, traditional project management techniques cannot meet the requirements for large-scale project scheduling. This dissertation is devoted to the optimization for the large-scale resource-constrained project scheduling problem based on a survey on the present research situation. Three stages are involved for capturing and solving the large-scale project scheduling problem. We firstly define the characteristics of large-scale projects which are the basis for mathematical model. Then this paper proposes a multi-objective optimization model, together with a solution algorithm in the second and third stage.To deal with defining the characteristics in the first stage, we analyze the research literatures in the last few years through a content analysis method. Based on the results of content analysis of large-scale project attributes, this paper focuses on the effects of stakeholders and risks on large-scale project scheduling. Since subcontracting is dominant in large-scale projects, the cost plus incentive fee contract is adopted to estimate the total contract cost of the general contractor. Meanwhile, the risk mitigation mechanism is adopted to quantify the impact of risks on the subproject’s makespan and cost. The costs of subcontracted large-scale projects are classified according to the project cost structure. Hence, a mathematical model for the large-scale resource-constrained project scheduling problem is established which aims to minimize the project makespan and total cost at the same time.To analyze the effectiveness of the proposed model, we design test sets for large scale problems and apply the NSGA-2algorithm to solve the problem. The result indicates that the model helps optimize the makespan and total cost of the large-scale project instances and that the NSGA-2is effective for large-scale resource-constrained project scheduling.