Resource and Space Constrained Project Scheduling
Open Access
- Author:
- Finke, Daniel
- Graduate Program:
- Industrial Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 16, 2010
- Committee Members:
- Deborah Jean Medeiros, Dissertation Advisor/Co-Advisor
Deborah Jean Medeiros, Committee Chair/Co-Chair
Mark Thomas Traband, Committee Member
Catherine Mary Harmonosky, Committee Member
John Messner, Committee Member
Michael Andrew Yukish, Committee Member
Vaughn Eugene Whisker Iii, Committee Member - Keywords:
- Congestion Modeling
4D Modeling
Project Scheduling
Genetic Algorithm - Abstract:
- Production planning in large-scale industries such as construction and shipbuilding is a key component in the manufacturing process and has a significant effect on the overall cost and schedule of the final product. This is due to the fact that the successful fabrication of these products requires a significant amount of material and resources, all of which must be coordinated and appropriately scheduled. Project scheduling methods have been used to model and schedule activities in these industries with great success. Until recently, spatial resources, such as floor space and interior space, have not been taken into consideration in the planning phase of a project. Traditionally, spatial conflicts have been resolved locally as they occur, often resulting in costly schedule delays. This thesis develops a framework to model space in a project scheduling problem comprised of installation activities within the interior of a large-scale product. The objective is to find a schedule with near-minimum makespan. The installation project activities are constrained by precedence relationships and limited resource availabilities. Space is modeled as a special resource. The space modeling approach approximates the space required by the activities and uses those space requirements to calculate spatial conflicts. Space conflicts are incorporated via a congestion function, which increases the duration of the installation activities with overlapping spatial requirements, reflecting productivity losses due to interference. The congestion function is based on the conflict volume and the length of time the conflict occurs. A prototype software application is developed from the framework to demonstrate the feasibility of the method. A hypothetical problem and a case study problem are solved using the application. The results show that considering spatial constraints increased the project makespan from 7% to 16% for the hypothetical problems and 35% for the case study problem. The developed framework establishes an integrated methodology to model the effects of worker and activity space congestion during schedule generation. In previous work, space conflicts were identified after the schedule was completed and resolved manually. The congestion function developed in this thesis explicitly considers types of space and the volume and time duration of conflicts among types. Previous work has either modeled multiple space types without estimating the effect on activity duration, or created a single function independent of space type. By incorporating realistic space requirements and considering the effect of spatial conflicts during schedule generation, a more realistic project schedule can be created that limits costly schedule delays attributed to less detailed planning practices. Congestion is inherently present in installation projects and has been managed by front line supervisors with little help from formal scheduling tools. The framework developed in this thesis enables front line supervisors to account for and control congestion to minimize schedule delays and project costs.