Auction implementation problems using LaGrangian Relaxation

LaGrangian Relaxation (LR) is used as an auction method for bidding in a de regulated environment. This paper focuses on the problems that occur while using LR as an auction method. Identical or similar units can prevent LR ho rn finding the optimal solution. If the units are similar, LR may have trou ble selecting some of them for the optimal solution. A feasible solution ma y thus not be found. Since a heuristic selection of which unit data to alte r has to be applied, this is inequitable to the unit(s)not selected and may result in lawsuit(s). Because the dispatcher has to use heuristic selectio n, there is no "fair'' solution to these problems. The authors suggest prev iously published methods to implement auctions. Sensitivity analysis has al so been applied in this paper to show the percentage difference of similari ty between units may affect the solution found.