Application of novel Nested decomposition techniques to long-term planning problems

Abstract

Cost effective, long term planning under uncertainty constitutes a significant challenge since a meaningful description of the planning problem is given by large Mixed Integer Linear Programming (MILP) models which may contain thousands of binary variables and millions of continuous variables. In this paper, a novel multistage decomposition scheme, based on Nested Benders decomposition is applied to the transmission planning problem. The difficulties in using temporal decomposition schemes in the context of planning problems due to the presence of non-sequential investment state equations are highlighted. An efficient and highly-generalizable framework for recasting the temporal constraints of such problems in a structure amenable to nested decomposition methods is presented. The proposed scheme's solution validity and substantial computational benefits are clearly demonstrated through the aid of case studies on the IEEE24-bus test system.