A Hybrid Neural Network Methodology for Cost Estimation

This paper presents an integrated systems approach, utilizing a hybrid methodology for the generation of practical cost and schedule baselines and for bid preparation under the prevailing competitive bidding environment. The system incorporates enhancements to the various functions that cover the quantitative aspects of estimate preparation, including: direct and indirect cost estimation, planning and scheduling, and resource utilization. For practicality, the system accounts for the qualitative aspects that strongly influence bid mark-up decision (e.g. Competition, market conditions, and contractor keenness for the job).In order to provide a practical and efficient aid to such an experience-based decision, the system utilizes Neural Networks, an AI-based technique that emulates the human ability to learn from past experience and derive speedy solutions to new situations, based mainly on intuitive judgement (gut feeling) that is difficult to model by other AI-based techniques including expert systems. The paper emphasizes the conceptual and design stages of the system development, establishing a structured methodology for cost estimation and bid preparation in a competitive environment.