Improved Lagrangian relaxation generation decision-support in presence of electric vehicles

Abstract

Decision making strategies for resources available in macro/micro scales have long been a critical argument. Among existing methods to address such a mixed-binary optimization model, lagrangian relaxation (LR) found universal acceptance by many utilities, offering a fast and accurate answer. This paper aims at retrofitting the solution way of LR algorithm by dint of meta-heuristic cuckoo search algorithm (CSA). When integrating CSA into LR mechanism, a tighter duality gap is catered, representing more accurate feasible solution. The key performance of CSA exhibits a head start over other classical methods such as gradient search (GS) and newton raphson (NR) when dealt with the relative duality gap closure in LR procedure. Further, electric vehicles (EV) with its associated hard constraints are encompassed into model to imperiling the proposed CSA-LR if encountered with nonlinear fluctuation of duality gap. Simulation results show that the proposed CSA-LR model outperforms the solution quality with/without EV as compared with conventional NR-LR method.