This thesis investigates the current approach to decision making in the construction industry with its background of uncertainty in relation to such factors as workload, production methods, resource availability and profitability. Given this situation, and the less than deterministic environment that usually surrounds the construction project, effective planning and control procedures are seen as a prime necessity and not as a luxury. Within this context, the most significant decisions relate to the time and cost parameters and more specifically to their interrelationship and the need to provide optimal or near optimal solutions to this relationship. A preliminary feasibility study was conducted in conjunction with six building construction companies operating in the United Kingdom. This indicated that the potential for substantial benefits exist and was further substantiated by replies received to a questionnaire circulated to one hundred additional construction companies. The time and cost parameters are investigated and the related decision problem formulated in a quantitative manner. Existing models are examined and three alternative models are postulated, viz. an integer linear programming model - this model, like the existing models studied, is difficult to implement due to the large number of variables and constraints involved; an integer linear programming model based on breakthrough points on the cost curve (since the number of breakthrough points is less than the number of points of discontinuity, there are fewer integer variables) and a heuristic model capable of dealing with the problems associated with nonlinear time cost curves using a microcomputer. Appropriate programs are developed for use on a CBM 32K microcomputer with a dual drive floppy disk system and high speed printer. Both the integer linear programming models and the heuristic model are tested using simulated project data. Comparative tests indicate that the heuristic model, although adopting a simpler method of analysis, is capable of providing a solution comparable in accuracy with the more sophisticated integer linear programming model. The computer system is designed to permit the data to be structured in several different ways depending on the needs of the recipient, ie., the person who makes the decisions receives only that part of the output that is relevant to their action.
|Date of Award