Smart management of electric vehicle integration: peak shaving and valley filling through locational marginal pricing and metaheuristic optimization

The integration of electric vehicles (EVs) with smart management systems is crucial for enhancing grid efficiency and stability. Techniques such as peak shaving and valley filling enable EVs to balance electricity demand, mitigate grid stress during peak hours, and utilize surplus energy during off-peak periods. This paper presents a centralized optimization algorithm for scheduling EV charging and discharging to improve grid load profiles through peak shaving and valley filling while satisfying grid and EV constraints. The optimization problem is solved using four metaheuristic algorithms: Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Differential Evolution (DE), and Cuckoo Search (CKS). The proposed framework, implemented in MATLAB, effectively fulfills the objective function, satisfies operational constraints, and minimizes adverse grid impacts. The algorithms are applied to 300 all-electric vehicles distributed across three charging stations. Incorporating locational marginal pricing (LMP) further enhances load profile management. A degradation-aware case study also integrates battery wear costs into the model. Results show that accounting for degradation leads EVs to adopt more conservative vehicle-to-grid (V2G) behavior, prioritizing battery health over short-term gains. However, targeted incentives during peak hours make discharging economically feasible, reducing peak loads and improving valley filling. This demonstrates that degradation-aware scheduling provides a more realistic and sustainable framework, while appropriate incentives can recover system-level benefits without compromising battery longevity. Simulation results indicate that the proposed method reduces peak and valley loads by 41.3%, with PSO achieving the fastest convergence, followed by DE, GA, and CKS.