Critical resource allocation in stochastic project networks

[No abstract. Extract from Introduction] This research is directed at an experimental investigation of the project management problem under conditions of limited resource availability; this problem is representative of a class of problems such as construction projects, ship repair, research projects, factory maintenance, and production scheduling. The problem is approached through the use of simulation and experimental methods in order to preserve the full richness of the problem. The ultimate objective is to provide the basis for a practical heuristic resource allocation procedure which derives its scheduling criteria from simulation of the project. Toward this end, the research attempts to provide insights into four interrelated aspects of the problem: 1. The introduction of stochastic activity durations, in order to observe the effects of this assumption on the measured project responses. 2. The characterization of the project by a set of variables, and the assessment of the effects of different levels of these variables on the measured project responses. 3. An evaluation of the relative performance of a number of currently used heuristic decision rules for resolving resource conflicts in constrained-resource projects. 4. The formulation of a more generalized project resource allocation procedure, in conjunction with the application of simulation methods, which will provide the project manager with a practical and useful management information system. To achieve these objectives, the research has proceeded through a series of logical steps, each involving an experiment, analysis of the data, and interpretation of the results. Chapter Two considers the broad aspects of the problem and sets the stage for later direct efforts to achieve the objectives. Specific areas of the project resource allocation problem are reviewed, compared, and summarized, leading to the formulation of the problem addressed by this research, in Chapter Three. Chapter Three establishes the direction of the research and defines the structure of the problem. The experimental vehicle is developed, transforming the problem concepts into a computer simulation model. The model is validated by a series of preliminary experiments. In Chapter Four, data from an initial screening factorial experiment is analyzed and interpreted, leading to a more comprehensive experiment in Chapter Five. Data from this experiment is extensively analyzed to illuminate the effects that the project variables have on the project duration and resource utilization. This analysis suggested the need for additional experimentation to establish the general features of a higher- order response surface. Chapter Five includes the design and analysis of a central composite experiment, and the interpretation of a quadratic response surface. A set of seven heuristic decision rules are compared over a wide range of project conditions in Chapter Six. Rules which provide consistently superior project schedules are identified, and insight is provided as to why different rules may be superior or inferior in project resource allocation. A search algorithm is combined with the project simulator to provide a means to find optimal allocation schedules in Chapter Seven. Cost-Duration Analysis and Critical Resource Allocation concepts are combined in the GERTS III-R simulator to demonstrate an application of the research findings. In Chapter Eight, the research is summarized, and further avenues for research are suggested.