Scholarlink

<b>Background</b> <p>The increasing deployment of electric vehicles in transportation is likely to experience rapid growth of electric vehicles (EVs). This requires efficient energy management during grid load variation to address large-scale integration and scheduling challenges and ensures cost-effectiveness, grid stability and battery longevity.</p> <b>Methodology</b> <p>To optimize the above challenges, the HOA is employed by adjusting the charging (G2V) and discharging (V2G) schedules. This paper presents a novel framework to improve dynamic scheduling accuracy and response. This framework integrated a hybrid model combining an artificial neural network (ANN) with the Hiking optimization algorithm (HOA). The proposed model is designed for optimal bidirectional power scheduling in the grid system assisted by photovoltaic (PV). The ANN provides the predictive foundation for HOA. The HOA utilizes and optimizes EV charging and discharging for improved load stability under various limitations. </p> <b>Results and discussion</b> <p>This hybrid ANN–HOA method demonstrates superior performance by leveraging learning analytics and adaptability for grid dynamics, achieving grid load balancing and reducing total operational costs from a baseline of ₹28,000 to ₹21,800. This provides a substantial 22% cost efficiency increase over conventional optimization approaches. Simulation results clearly represent that the proposed HOA achieves faster convergence speed, lower peak grid load and improved economic efficiency compared to Particle Swarm Optimization (PSO), Grey Wolf Optimization (GWO) and Genetic Algorithm (GA). Moreover the proposed model is evaluated under varying grid load targets (650–750 kW) and initial EV SoC levels (40–90%). The proposed algorithm gives the result of faster convergence with the iteration of 8 and the fitness value of 0.122. </p> <b>Conclusion</b> <p>This ANN–HOA model emerges as a highly effective and reliable solution for real-time load optimization and cost-effective EV scheduling.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Hybrid Artificial Neural Network-based Hiking Optimization in Optimal EV Scheduling for Vehicle to Grid and Grid to Vehicle

  • 期刊论文

  • 发表日期：

    2025年12月4日

查看全文
  • Nageswara Rao,
  • Shelly Vadhera,
  • Shashi Bhushan Singh

摘要

Background

The increasing deployment of electric vehicles in transportation is likely to experience rapid growth of electric vehicles (EVs). This requires efficient energy management during grid load variation to address large-scale integration and scheduling challenges and ensures cost-effectiveness, grid stability and battery longevity.

Methodology

To optimize the above challenges, the HOA is employed by adjusting the charging (G2V) and discharging (V2G) schedules. This paper presents a novel framework to improve dynamic scheduling accuracy and response. This framework integrated a hybrid model combining an artificial neural network (ANN) with the Hiking optimization algorithm (HOA). The proposed model is designed for optimal bidirectional power scheduling in the grid system assisted by photovoltaic (PV). The ANN provides the predictive foundation for HOA. The HOA utilizes and optimizes EV charging and discharging for improved load stability under various limitations.

Results and discussion

This hybrid ANN–HOA method demonstrates superior performance by leveraging learning analytics and adaptability for grid dynamics, achieving grid load balancing and reducing total operational costs from a baseline of ₹28,000 to ₹21,800. This provides a substantial 22% cost efficiency increase over conventional optimization approaches. Simulation results clearly represent that the proposed HOA achieves faster convergence speed, lower peak grid load and improved economic efficiency compared to Particle Swarm Optimization (PSO), Grey Wolf Optimization (GWO) and Genetic Algorithm (GA). Moreover the proposed model is evaluated under varying grid load targets (650–750 kW) and initial EV SoC levels (40–90%). The proposed algorithm gives the result of faster convergence with the iteration of 8 and the fitness value of 0.122.

Conclusion

This ANN–HOA model emerges as a highly effective and reliable solution for real-time load optimization and cost-effective EV scheduling.