Abstract
Competitive bidding situations involve considering a multiplicity of factors. Organizations must be able to weigh the relative probability of potential projects based on resource usage, project duration and competitor actions to decide which of many possible bids to submit. A bidding strategy designed to maximize expected long run return is crucial, since an organization can usually submit only one bid per project.
This paper presents a family of stochastic dynamic programming models considering different bidding situations. Several projects, each with several potential bids, are available for each situation. The objective is to determine what bidding strategy will maximize expected returns. Models are developed for two principle bidding situations: sequential, where projects are bid individually; and simultaneous, where several projects are bid at one time. Next, the effects of over- or under-commitment of resources are incorporated into the models. Finally, changes in project timing and the resultant effects on bidding strategy are included.
A numerical example traces the changes in bidding strategy which occur as the models are expanded. The general formulation of bidding problems is also discussed, including changing the bid success probabilities due to competitor actions, the possibility of crashing projects and alternate methods of performing projects.