Embedding sensitivity theory in ordinal optimization for decentralized optimal power flow control

In this paper, the decentralized optimal power flow with continuous and discrete control variables problem is firstly formulated as a NP-hard optimization problem - Block Additive constrained with Continuous and Discrete variables (BACD) problem. Secondly, an algorithm of embedding sensitivity theory (ST) in ordinal optimization (OO), abbreviated as STOO, is proposed for solving this NP-hard optimization problem. The STOO algorithm consists of three stages and three models of performance evaluation. The proposed method not only copes with the computational complexity due to huge solution space but also obtains a good enough solution with high probability guaranteed by the OO theory. Finally, this work demonstrates the computational efficiency of the STOO algorithm via various tests on the IEEE 118-bus and 244-bus systems partitioned into four subsystems using a 4-PC network and compares the results with those obtained using other heuristic methods, Genetic Algorithm, Tabu Search, Ant Colony Optimization and Simulated Annealing. Test results show the validity, robustness and excellent computational efficiency of the STOO algorithm for obtaining a good enough solution.

► Decentralized optimal power flow with continuous and discrete variables.
► Algorithm of embedding sensitivity theory in ordinal optimization.
► Block additive constrained with continuous and discrete variables.
► IEEE 118-bus and 244-bus systems partitioned into four subsystems.