Cost Predictions of Facilities

Abstract

Facility managers in all organizations experience many different problems. One of the most common significant problems is the prediction of future operating and maintenance costs. Although extremely accurate cost predictions are difficult, information generated through historical cost analyses, along with a study of the factors that determine these costs, could contribute greatly to improving the accuracy of such predictions. In the research described, 14 university facilities and eight government offices were investigated. Quantitative data on the historical operating and maintenance costs of these facilities, along with knowledge of the factors affecting the costs were elicited through various sources. Cost prediction models were developed using neural networks, regression analyses, and random deviation detection methods. The final stage of the research was the creation of a costs prediction decision-support system using the analytical results and the knowledge acquired. The system created may be used to assist and advise on certain aspects of facility management, such as the estimation of operating and maintenance costs, and the development of preventive and general maintenance plans for facilities similar to those investigated.