The highs and lows of unbalanced bidding models

The unbalanced bidding models developed in the first 50 years, since Marvin Gates first invented them in 1956, have suffered from a significant common flaw. Typically designed as linear programming models, with the objective being to maximise the contractor’s profitsfrom a project, they have incorporated constraints on the prices for each of the items such that they are each bound by lower and upper limits. The intent of this was to find optimum prices falling somewhere within these limits. Instead, the effect of these models has beenthat all optimal prices (barring only one) are found to lie exactly on the extreme edge of these limits. In effect then, these models serve only to decide which items should beassigned their lowest acceptable price, and which items should be assigned their highest acceptable price. Tests done on a series of simulated hypothetical projects, created randomly by way of an automated process, illustrate this effect, which has previously not being observed. This effect is suggested as being undesirable – these pricing boundaries are vague and heuristically difficult to determine and hence relatively ‘soft’ in nature, ratherthan being inelastic and hard-and-fast. The risks - that these limits are designed to avoid – are not of the nature that they are incurred (fully) marginally beyond these limits and yet not incurred at all within the limits. Nevertheless, even though these boundaries are only vaguely definable by nature, these models do somehow need to acknowledge that extreme prices are unacceptable and normal (‘central’) prices are fine. This problem has been solvedwith the use of component unit pricing (CUP) theory.