Hybrid Dynamic Programming/Branch-and-Bound Strategies for Electric Power Generation Planning

Abstract

This paper espouses the application of hybrid dynamic programming/branch-and-bound theory to the generation planning problem (GPP) of electric utility companies. A hybrid approach retains the advantages of previous dynamic programming formulations of the GPP; these advantages include the relative ease with which the more sophisticated and realistic measures of the system's reliability and variable operating cost can be represented within the model framework. Meanwhile, the disadvantages of pure dynamic programming, including the relatively large amounts of CPU time and high speed storage space required in solving GPP's, can be ameliorated by a hybrid approach. This reduction in CPU time and high speed storage space requirements could allow electric utilities to solve larger GPP's, thereby permitting planners to study problems with longer planning horizons and a greater number of feasible solutions.