Job description

Joining LINEACT at CESI for a research internship would be a fantastic opportunity to contribute to innovative projects while deepening my skills in a cutting-edge environment focused on digital transformation and Industry 4.0.

Abstract

Keywords: Optimization, Electric Mobility, Multi-objective Optimization, Online Algorithms, Real-Time Optimization.

The global transition to electric vehicles (EVs) is accelerating, driving the need for efficient and scalable charging infrastructure. Optimizing resource allocation in EV charging stations is needed to meet this growing demand effectively. This internship aims to develop scheduling strategies for charging stations with limited resources while managing the uncertainties of EV arrival and departure times. The objective is to maximize the number of fulfilled charging demands, subject to constraints such as power capacity and charger availability. To tackle uncertainty, online algorithms and reinforcement learning techniques will be explored to enable dynamic, data-driven decision-making. While previous studies have focused on offline scheduling with predefined parameters, real-world applications demand adaptive solutions to accommodate fluctuating demand patterns. This project will leverage real-time data to enhance charging schedules and prevent grid overload. Smart algorithms will be used to optimize charging priorities based on energy costs, grid load, and user preferences. Furthermore, multi-objective optimization methods will be applied to balance two competing objectives: reducing charging costs under time-of-use electricity tariffs while maximizing service availability. By implementing these strategies, the project aims to create a sustainable, user-friendly, and efficient EV charging infrastructure, contributing to the broader goal of future-ready mobility solutions.

Research Work

Scientific context

The global transition toward EVs is accelerating, driven by environmental concerns and government policies to reduce greenhouse gas emissions. Global electric car sales are expected to surpass 17 million units in 2024, bringing the total number of EVs on the road to 40 million [1]. Naturally, these vehicles will require widespread charging infrastructure, much like traditional internal combustion engine (ICE) vehicles, which depend on fuel stations for refueling. However, unlike ICE vehicles, which can refuel in just a few minutes, EVs generally take much longer to charge. For example, a typical Level 2 home charger can take around 4 to 10 hours for a full charge, while a DC fast charger can reduce that time to 30 minutes to an hour, depending on the battery size (ref).

EV charging can be done at private locations, such as home garages or public charging stations, most likely for long-distance travel and urban areas. While home charging remains the most common method for recharging EVs today, it is not a universal solution for all. In cities with high population density, many live in multi-unit dwellings where access to home charging is often limited. This restriction forces EV owners to rely more on public charging infrastructure. In a recent survey of European EV drivers conducted in 2023 [2], 44% of respondents did not have a home charging point installed. Similarly, a U.S. study ranked insufficient public charging stations as the second most significant barrier to EV adoption [3]. Globally, inadequate charging infrastructure continues to be one of the primary obstacles, limiting flexibility and comfort for users and hindering widespread EV adoption.    

Despite these challenges, significant efforts are underway to deploy more charging points and public charging stations. The development and interoperability of public charging networks continues to accelerate. In 2023, the public charging stock grew by over 40%, with the expansion of fast chargers—up by 55%—outpacing the growth of slower charging options. This trend highlights the critical role of public charging infrastructure in supporting the continued growth of the EV market.

Despite the positive trend in the growth of public charging stations, challenges still loom on the horizon. One primary concern is the potential strain on the power grid as the number of EVs rises. Overloading the grid could become a serious issue, especially during peak charging times. Additionally, as the demand for charging points grows, finding affordable and accessible public stations may become more complicated and burdensome for EV owners.

Charging at a slow or fast speed is another important consideration for consumers looking to switch to EVs, particularly for long journeys. Consumers expect charging services to be fast, easy to use, reliable, and transparently priced. To mitigate these challenges, it will be essential for EV owners to have access to real-time charging information. This capability will allow them to locate available charging points upon arrival at their destination—whether at work, in a building's parking lot, or elsewhere—and to charge their vehicles to the desired state of charge. Real-time access to charging data will reduce the stress of finding charging opportunities and ensure a seamless EV ownership experience.

In this internship, we studied innovative charging approaches integrating real-time algorithms to optimize charging times and grid load. Smart charging helps distribute energy demand evenly across the grid and allows EVs to charge during off-peak hours when energy is cheaper and more readily available. These algorithms can prioritize charging based on energy costs, grid load, and EV owner preferences. Using real-time data, charging stations can dynamically adjust charging speeds, ensuring the grid is not overloaded while meeting user needs. By implementing smart charging strategies, we can create a more efficient and sustainable charging infrastructure that benefits the grid and EV owners.

Subject

In this internship, we will optimize EV charging strategies in stations with limited resources under uncertainty. Specifically, we will address the challenge of scheduling EV charging in a station with limited overall power capacity and a restricted number of chargers. When EV drivers submit their charging demands, the scheduler must decide, given these constraints, whether to accept or reject each demand. Accepted demands must then be fully satisfied within the available time window. The primary objective of the scheduler is to maximize the number of satisfied demands while managing the station’s limited resources efficiently.

A key complexity of this problem is the uncertainty surrounding the arrival and departure times of the EVs. Drivers may not arrive or depart precisely when expected, adding more difficulty to the scheduling process. To tackle this, we will explore online algorithms that adapt to real-time changes in the system and employ reinforcement learning techniques to make dynamic, data-driven decisions. These approaches allow the scheduler to react to evolving conditions, optimizing resource use as new information becomes available.

Previous research [5, 6, 7] has explored the offline version of this problem, assuming complete knowledge of EV arrivals and departures in advance. Various theoretical results have been established, and several solution approaches, including linear programming models, heuristic methods, and metaheuristics, have been proposed to achieve high-quality solutions with reduced computational time. However, real-world applications require an online approach to handle the inherent uncertainty in EV charging demand. This internship will focus on developing such an approach to enhance adaptability and efficiency in dynamic environments.

Furthermore, multi-objective optimization methods have already been studied for planning problems [8, 9, 10, 11]. In this internship, we will address the complex challenges of managing EV charging infrastructures. These methods will balance between two critical and often competing objectives: minimizing charging costs under dynamic time-of-use electricity tariffs, which fluctuate based on grid demand and energy availability, and maximizing service availability to ensure that charging stations are accessible and operational when and where users need them. This internship aims to develop an EV charging infrastructure energy manager that is efficient, scalable, and sustainable. The proposed strategies will enhance operational performance, reduce costs, and support environmental goals. Optimizing energy use will contribute to the widespread adoption of EVs and address key challenges in urban mobility.

Prior works in the laboratory

CESI LINEACT focuses on sustainable cities by conducting research on eco-friendly transport systems, including initiatives such as "Mon trajet vert" [4]. The laboratory has also explored electric vehicle charging solutions [5, 6, 7] and shared micro-mobility, incorporating prediction techniques to optimize operations. Additionally, CESI LINEACT addresses planning and scheduling challenges in various industrial processes, considering complex problem variants such as multi-objective optimization [8, 9, 10, 11] and real-time decision-making scenarios.

Work program

Below is a detailed outline of this internship:  

·       Review the existing studies on online EV charging scheduling problems.

·       Explore reinforcement learning (RL) techniques for optimizing charging strategies.

·       Investigate multi-objective optimization techniques to balance competing goals such as cost, energy efficiency, and user satisfaction.

·       Design the algorithms (heuristic, RL-based, and multi-objective) for the EV charging scheduling problem.

·       Implement algorithms within the EV charging station simulation environment to test and validate their performance.

·       Test algorithms on various charging demand scenarios (e.g., peak vs. off-peak demand, varying levels of uncertainty).

·       Explore the impact of different levels of uncertainty in arrival/departure times on scheduling effectiveness.

·       Compare offline and online approaches.

Expected scientific/technical production 

·       Final Report: Comprehensive report detailing the methodology, simulation results, and algorithm performance.

·       Code and Simulations: Well-documented code implementing the proposed algorithms.

·       Conference Paper: Presentation of findings at a relevant academic conference. This could include poster presentations or full papers, providing opportunities to discuss the research with other experts in the field and receive valuable feedback.

Context

Lab presentation

CESI LINEACT (UR 7527), Laboratory for Digital Innovation for Businesses and Learning to Support the Competitiveness of Territories, anticipates and accompanies the technological mutations of sectors and services related to industry and construction. The historical proximity of CESI with companies is a determining element for our research activities. It has led us to focus our efforts on applied research close to companies and in partnership with them. A human-centered approach coupled with the use of technologies, as well as territorial networking and links with training, have enabled the construction of cross-cutting research; it puts humans, their needs and their uses, at the center of its issues and addresses the technological angle through these contributions.

Its research is organized according to two interdisciplinary scientific teams and several application areas.

-      Team 1 "Learning and Innovating" mainly concerns Cognitive Sciences, Social Sciences and Management Sciences, Training Techniques and those of Innovation. The main scientific objectives are the understanding of the effects of the environment, and more particularly of situations instrumented by technical objects (platforms, prototyping workshops, immersive systems...) on learning, creativity and innovation processes.

-      Team 2 "Engineering and Digital Tools" mainly concerns Digital Sciences and Engineering. The main scientific objectives focus on modeling, simulation, optimization and data analysis of cyber physical systems. Research work also focuses on decision support tools and on the study of human-system interactions in particular through digital twins coupled with virtual or augmented environments.

These two teams develop and cross their research in application areas such as

-      Industry 5.0,

-      Construction 4.0 and Sustainable City,

-      Digital Services.

Areas supported by research platforms, mainly that in Rouen dedicated to Factory 5.0 and those in Nanterre dedicated to Factory 5.0 and Construction 4.0.

Links to the research axes of the research team involved

The internship topic aligns with the research activities of the Engineering and Digital Tools research team of CESI LINEACT, particularly within Multimodal transport systems management thematic.

Desired profile

Required skills

Scientific and technical skills: The selected candidate must:

·       Be in the final year of a master’s degree program in engineering, operations research, computer science, applied mathematics, or any related field.

·       Strong knowledge of optimization algorithms and reinforcement learning.

·       Strong programming skills in computer languages: C/C++/Python.

·       Work on GitHub.

Soft skills: The selected candidate must:

·       Be autonomous, have initiative and curiosity.

·       Know how to work in a team and have good interpersonal skills.

·       Be rigorous.

·       Can analyze problems and data systematically.

·       Can clearly and effectively communicate ideas.

·       Uphold strong ethical standards in research.

Bonus at 15% of the Social Security hourly ceiling.

Starting date: February 2025

Bibliography:

1.     International Energy Agency. Global EV Outlook 2024, 2024. Licence : CC BY 4.0.

2.     Shell Recharge. Ev driver survey report 2023, June 2023. In partnership with LCP Delta.

3.     Pamidimukkala, A., Kermanshachi, S., Rosenberger, J. M., & Hladik, G. (2023). Evaluation of barriers to electric vehicle adoption: A study of technological, environmental, financial, and infrastructure factors. Transportation Research Interdisciplinary Perspectives, 22, 100962.

4.     https://www.montrajetvert.fr/

5.     Zaidi, I., Oulamara, A., Idoumghar, L., & Basset, M. (2024). Minimizing grid capacity in preemptive electric vehicle charging orchestration: Complexity, exact and heuristic approaches. European Journal of Operational Research, 312(1), 22-37.

6.     Zaidi, I., Oulamara, A., Idoumghar, L., & Basset, M. (2023). Electric vehicle charging scheduling problem: Heuristics and metaheuristic approaches. SN Computer Science, 4(3), 283.

7.     Zaidi, I., Oulamara, A., Idoumghar, L., & Basset, M. (2024). Maximizing the number of satisfied charging demands of electric vehicles on identical chargers. Omega, 127, 103106.

8.     Belhocine, L., Dahane, M., & Yagouni, M. (2021). Customer behaviour–based multi-objective approach for the recovery and remanufacturing of used products: application to smartphone reconditioning process. The International Journal of Advanced Manufacturing Technology, 117, 125-146.

9.     Belhocine, L., Dahane, M., & Yagouni, M. (2020, October). Heuristic based strategy for multi-components products recovery and remanufacturing. In 2020 5th International Conference on Logistics Operations Management (GOL) (pp. 1-6). IEEE.

10.  Belhocine, L., Dahane, M., & Yagouni, M. (2020, June). Performance and usage frequency-based policy for products recovery and remanufacturing. In 2020 7th International conference on control, decision, and information technologies (CoDIT) (Vol. 1, pp. 849-854). IEEE.

11.  Belhocine, L., Dahane, M., & Yagouni, M. (2019, October). Contribution to the optimisation of products recovery and remanufacturing: a multiobjective non-dominated sorting genetic algorithm-based hybrid approach. In 2019 IEEE International Conference on Systems, Man and Cybernetics (SMC) (pp. 134-139). IEEE.

#CESILINEACT