Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LLM-Enabled EV Charging Stations Recommendation

Zeinab Teimoori

TL;DR

The paper tackles the slow expansion of EV charging infrastructure and fragmented CS data by introducing RecomBot, a prompt-based, LLM-enabled recommender that fuses real-time multi-modal data. It leverages Retrieval-Augmented Generation (RAG), user-preference normalization, and a constrained optimization framework to produce ranked CS recommendations, with personalization enhanced through fine-tuning on EV-specific data and a reinforcement-learning feedback loop. Key contributions include a 5-step RecomBot pipeline, real-time API integration (e.g., Open Charge Map, Google Cloud), and demonstration of adaptive ranking across diverse user prompts. The evaluation in Kamloops shows that RecomBot can effectively retrieve nearby stations and tailor suggestions to user preferences, suggesting practical improvements in charging efficiency and user experience, with future work expanding to mobile CSs and environmental factors.

Abstract

Charging infrastructure is not expanding quickly enough to accommodate the increasing usage of Electric Vehicles (EVs). For this reason, EV owners experience extended waiting periods, range anxiety, and overall dissatisfaction. Challenges, such as fragmented data and the complexity of integrating factors like location, energy pricing, and user preferences, make the current recommendation systems ineffective. To overcome these limitations, we propose RecomBot, which is a Large Language Model (LLM)-powered prompt-based recommender system that dynamically suggests optimal Charging Stations (CSs) using real-time heterogeneous data. By leveraging natural language reasoning and fine-tuning EV-specific datasets, RecomBot enhances personalization, improves charging efficiency, and adapts to various EV types, offering a scalable solution for intelligent EV recommendation systems. Through testing across various prompt engineering scenarios, the results obtained underline the capability and efficiency of the proposed model.

LLM-Enabled EV Charging Stations Recommendation

TL;DR

The paper tackles the slow expansion of EV charging infrastructure and fragmented CS data by introducing RecomBot, a prompt-based, LLM-enabled recommender that fuses real-time multi-modal data. It leverages Retrieval-Augmented Generation (RAG), user-preference normalization, and a constrained optimization framework to produce ranked CS recommendations, with personalization enhanced through fine-tuning on EV-specific data and a reinforcement-learning feedback loop. Key contributions include a 5-step RecomBot pipeline, real-time API integration (e.g., Open Charge Map, Google Cloud), and demonstration of adaptive ranking across diverse user prompts. The evaluation in Kamloops shows that RecomBot can effectively retrieve nearby stations and tailor suggestions to user preferences, suggesting practical improvements in charging efficiency and user experience, with future work expanding to mobile CSs and environmental factors.

Abstract

Charging infrastructure is not expanding quickly enough to accommodate the increasing usage of Electric Vehicles (EVs). For this reason, EV owners experience extended waiting periods, range anxiety, and overall dissatisfaction. Challenges, such as fragmented data and the complexity of integrating factors like location, energy pricing, and user preferences, make the current recommendation systems ineffective. To overcome these limitations, we propose RecomBot, which is a Large Language Model (LLM)-powered prompt-based recommender system that dynamically suggests optimal Charging Stations (CSs) using real-time heterogeneous data. By leveraging natural language reasoning and fine-tuning EV-specific datasets, RecomBot enhances personalization, improves charging efficiency, and adapts to various EV types, offering a scalable solution for intelligent EV recommendation systems. Through testing across various prompt engineering scenarios, the results obtained underline the capability and efficiency of the proposed model.
Paper Structure (4 sections, 9 equations, 4 figures, 2 tables)

This paper contains 4 sections, 9 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Evaluation and Results
  4. Conclusion and Future Work

Figures (4)

  • Figure 1: Insufficient growth of EV charging infrastructure vs. EV sales EV_Outlook.
  • Figure 2: RecomBot agent framework integrates prompt engineering and optimization techniques to refine selections based on user preferences.
  • Figure 3: Station retrieved from the API surrounding the EV user location (in blue).
  • Figure 4: RecomBot performance based on the user preference "fast charging high rated".