A general approach for solving the Resource-Constrained Project and Multi-Project Scheduling Problem
Date
2020-11-02
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Abstract
The project scheduling problem belongs to the complex decision-making process, which has been increasingly attracting attention from both researchers and practitioners. Researchers have studied resource-constrained project scheduling problems (RCPSPs) and resource-constrained multi-project scheduling problems (RCMPSPs), focusing on limited or even inadequate resources as constraints to implementing project activities and decreasing completion times. These constraints typically make project scheduling problems highly challenging to deal with. Most of the current project scheduling software packages and scheduling techniques, which assume that activities, once started, cannot be interrupted. In other words, each project activity is pre-planned and cannot be re-scheduled into two or more sub-activities. Project activities can be interrupted due to limited resources and according to case studies in this thesis, it validates that this assumption leads to significant improvements in project schedules. Activities under traditional assumptions (i.e., under category A) can be performed using fixed resources along the Y-axis (resource axis) over a fixed duration along the X-axis (time axis) and cannot be interrupted. In this thesis, the traditional RCPSP is referred to as RCPSP(A). However, in practice, some project activities can be interrupted during the project execution. Consequently, the RCPSP should consider not only project activities under category A but also two more categories: B and D. Category B applies to activities that use fixed resources along the Y-axis but can be interrupted along the X-axis. Category D refers to activities that can be flexibly re-scheduled along both the Y-axis (changeable resources) and X-axis (interrupted or not interrupted). Under the classification of project activities into categories A, B and D individually, the project scheduling problems are correspondingly referred to as RCPSP(A), RCPSP(B) or RCPSP(D), respectively. Many of the existing papers and commercially available project scheduling software packages deal with project scheduling problems that belong to RCPSP(A); however, the RCPSP(B) and RCPSP(D) are rarely addressed. In the RCPSP(A) and RCPSP(B), project activity is pre-planned in advance by a fixed resource requirement over a total time duration, e.g. 4 men over 6 days, whereas, in the RCPSP(D), the activities are pre-planned by a fixed work content, e.g. 24 man-days. However, A, B and D categories can exist simultaneously in engineering projects, such as a one-of-a-kind production (OKP) project for making customized oil cargo, which leads the RCPSP to generally include the three categories simultaneously, simply recorded as RCPSP(ABD). The RCPSP(ABD) has not been considered in existing papers and commercially available project scheduling software packages. In fact, the RCPSP(ABD) is a more general view of RCPSP, and the RCPSP(A), RCPSP(B) or RCPSP(D) can be viewed as a special case of the RCPSP(ABD). In make-to-order and the Design-to-order manufacturing companies, to better focus on customer demands and specific requirements, each customer order is treated as a project in the company, and the company has to handle multi-projects to meet all its customers’ demands. Project activities are classified under category A, and the production schedule is subject to three constraints: 1) resources are limited, 2) customer requests are considered as projects, and 3) projects are initiated based on demand time. The production schedule in these types of production companies can be considered as the resource-constrained multi-project scheduling problem under activity categories A, B and D simultaneously (RCMPSP(ABD)) in an integrated manner. To improve production efficiency and reduce costs, the resources must be shared among the projects. Thus, an effective scheduling algorithm is needed to integrate all project schedules into one production plan and then to optimize this plan, i.e. to find the shortest makespan for the production. In this thesis, the research focus is placed on the RCPSP(ABD) /RCMPSP(ABD), and a new priority rule-based heuristic has been developed to examine the impact of the interruptible and changeable activities assumptions in terms of shortening the project or multi-project makespan and increasing the average utilization rate of resources. The thesis finds alternative schedules and helps project managers selectively choose how specific activities are to be planned and re-scheduled in order to improve project scheduling efficiency. The heuristic identifies the critical interruptible activities that have significant impacts on the project makespan and facilitates project activity planning and re-scheduling under categories A, B and D simultaneously.
Description
Keywords
Citation
Ben Issa, S. (2020). A general approach for solving the Resource-Constrained Project and Multi-Project Scheduling Problem (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.