Multi-project scheduling under multiple resource constraints: an experimental investigation

Abstract

The term Multi-Project Scheduling under Multiple Resource Constraints, MPSMRC, refers to a situation where an organization seeks to allocate a set of common scarce resources to the competing activities of a number of different projects on hand. Usually, heuristic decision rules are resorted to prioritize the competing activities and subsequently, based on these priorities, allocate the scarce resources. The present work aims at a comprehensive study of the MPSMRC problem. It takes a look at the past approaches and makes an assessment of their areas of strengths and weakness-soon subsequently, a scheduling system is developed which is based upon such or newly developed better procedures of prioritization. In the development of such a scheduling system, it is necessary on our part to be able to predetermine certain criteria for the evaluation of schedules. Based on the field studies as well as studies by means of limited surveys, the following measures were selected for the evaluation of Schedules: (i) Minimization of total project set delays. (ii) Minimization of the Maximum individual project delay. The study was carried out for the following two cases, (1) Activity time durations are known with certainty (Deterministic case) and, (ii) Activity time flow known probability distributions (stochastic case}. A two-factor experimental design was used for the study; Computer-based experiments were used to evaluate the heuristic rules developed in the course of this work The major contribution of the study is the development of a decision framework, based on the heuristic rules developed for the solution of this problem. This can help the project managers in taking Resource Allocation decisions in a dynamic Multi-Project environment. The other associated findings and contributions of this study are mentioned below: (i) The development of adaptive rules & new composite rules for multi—project scheduling; (ii) The establishment of Activity duration, Activity slack, and Project due-date as the key parameters for evolving a good scheduling rule; (iii) The recognition of the importance of Resource Availability to Resource Requirement Ratio, henceforth termed as Resource Availability Factor (RAF), as a key factor in choosing an appropriate heuristic decision rule to resolve resource conflicts; and (iv)It was also found that the consideration of uncertainty of activity durations did not significantly alter the relative performance of rules identifier for the deterministic activity case.