Skip to Main Content
This article proposes the use of Lagrange multipliers associated with discrete variables of the Thermal Unit Commitment problem as a source of information for the ant colony algorithm. To achieve this, the discrete variables that are inherent to the problem are mitigated through a sigmoid function. By doing so, the non-linear optimization issue is solved through the use of the primal-dual interior-point method, generating Lagrange multipliers associated to the ON/OFF decision variables as subproducts which are used to draw up a list of priorities, where part of the colony will make use of this information in the search for solutions. The results obtained show that the information taken into consideration significantly improves the efficiency of the colony search process.
PowerTech, 2011 IEEE Trondheim
Date of Conference: 19-23 June 2011