Double spatial scale search algorithm for distributed multi-area economic dispatch via problem pre-processing

Distributed multi-area economic dispatch (MAED) can overcome the deficiencies of a large amount of information processing and regional privacy leakage faced by centralized scheduling. However, multiple iterations may increase communication and computing costs, as well as operational risk. Therefore, this paper proposes a double spatial scale search (DSSS) algorithm to accelerate the iterative process. First, each area performs the multi-parametric quadratic programming (MPQP) with load and phase angles of boundary buses as parameters to pre-process the MAED problem as a piecewise quadratic programming (PQP) problem. To reduce the complexity of MPQP, a parameter variables reduction method and a fitting-based imprecise MPQP (i-MPQP) algorithm are developed. Then, aided by the i-MPQP results, we design a DSSS algorithm to optimize the PQP problem efficiently. Finally, two interconnected power systems of different sizes are used for numerical testing. The results show that the i-MPQP can produce a 77.98% and 88.65% reduction in the number of precise MPQP (p-MPQP) function segments. Compared with the existing algorithm, the iterations of DSSS algorithm are decreased by 75% and 76.35%, and it can converge with one iteration in most cases. Moreover, the performance of the proposed method is insensitive to the system settings and algorithm parameters.