Development of a decision support system (DSS) for the contractor's decision to bid: regression and neural networks solutions

The decision whether to bid or not for a project is extremely important to construction contractors; besides the issues of resource allocation, the preparation of a bona fide tender commits the organisation to considerable expenditure, which is only recovered if the bid is successful. There is, therefore, a potential financial benefit to be realised through the adoption of an effective and systematic approach to the decision to bid process. Artificial neural network and regression techniques are used to model data collected from the bid/no-bid decision makers of a UK construction company for 115 historical bid opportunities. While the regression model is ultimately rejected, the selected back-propagation network, comprising 21 input nodes, 3 hidden layers and 4 output nodes is used to support a DSS for the decision to bid process. The results obtained demonstrate that the model functions effectively in predicting the decision process.