Bidding information to generate bidding strategies for LaGrangian relaxation-based auctions

LaGrangian relaxation (LR) is used as an auction method for bidding in a de regulated environment. Identical or similar units can prevent LR from findi ng the optimal solution when only one of the units should be committed. If many units are similar, LR may have trouble selecting some subset of them f or the optimal solution. A unique feasible solution may thus not be found. This leads to inequity among the unit(s) not selected and may result in les s revenue for one or more competitors. Because the dispatcher has to use he uristic selection, there is no 'fair' solution to these problems. This pape r focuses on how to change unit data to obtain an advantage while using LR as an auction method. The authors suggest alternative strategies based on p reviously published problems with selection by unit commitment and subseque nt dispatch by economics. Sensitivity analysis results demonstrate the meth od for finding the percentage difference between units to affect the soluti on. (C) 1999 Elsevier Science S.A. All rights reserved.