Failure localization in time critical market applications

Abstract

In time critical market applications such as for example scheduling and price computation for the balancing market, failure of the algorithm in finding a solution would result in cancelation of the session and respective financial consequences for the market participants. In the paper we propose a regularization procedure that could help an operator to spot the problematic location and find out the reasons that caused the algorithm to breakdown and make necessary corrections. The approach has proved to be useful in the operation of the Balancing Market in the Russian Federation.