Optimal scheduling of integrated energy system using decoupled distributed CSO with opposition-based learning and neighborhood re-dispatch strategy

Integrated energy optimization scheduling (IEOS) is a complex problem aiming to minimize the total cost while the requirements of load balance is met. Due to the non-convex, non-differentiable and high-dimensional characteristics, there are many difficulties in solving the problem. Based on a regional integrated energy system (RIES), a decoupled distributed crisscross optimization with opposition-based learning and neighborhood re-dispatch strategy (DDCSO-OBL-NR) is proposed to solve IEOS problem by distributed method with different energy types as the scale. Initially, the CSO with excellent global search ability is firstly used to solve the complicated IEOS problem. Then, based on the distributed structure, distributed parallel computing can be achieved by DDCSO, which contributes to 1) protect the privacy of different energy data, 2) reduce the solving dimensions and 3) relieve the heavy communication burden. The total optimal cost is achieved by minimizing the cost of each portion without centralized controller. Furthermore, the opposition-based learning (OBL) strategy and the neighborhood re-dispatch (NR) strategy are combined into DDCSO aiming to optimize initial population location and enhance local search ability. Eventually, the DDCSO-OBL-NR is realized, and the effectiveness of which in solving the distributed IEOS problems is verified by the experimental results of three cases.