The growing demand for electric vehicle (EV) charging infrastructure presents significant planning challenges, requiring efficient strategies for investment and operation to deliver cost-effective charging services. However, the potential benefits of EV charging assignment, particularly in response to varying spatial-temporal patterns of charging demand, remain under-explored in infrastructure planning. This paper proposes an integrated approach that jointly optimizes investment decisions and charging assignments while accounting for spatial-temporal demand dynamics and their interdependencies. To support efficient model development, we leverage a large language model (LLM) to assist in generating and refining the mathematical formulation from structured natural-language descriptions, significantly reducing the modeling burden. The resulting optimization model enables optimal joint decision-making for investment and operation. Additionally, we propose a distributed optimization algorithm based on the Alternating Direction Method of Multipliers (ADMM) to address computational complexity in high-dimensional scenarios, which can be executed on standard computing platforms. We validate our approach through a case study using 1.5 million real-world travel records from Chengdu, China, demonstrating a 30% reduction in total cost compared to a baseline without EV assignment.
Electric vehicles (EVs) have emerged as a promising solution to reduce greenhouse gas emissions in the transportation sector, thereby advancing the goal of net-zero transportation. The adoption of EVs can reduce operational and maintenance costs for consumers but offer significant environmental benefits arXiv:2511.19055v1 [eess.SY] 24 Nov 2025 for society [1]. Research indicates that widespread EV adoption could mitigate up to 25.5% of anticipated winter warming [2]. Leading this transformative change, countries such as the United Kingdom are actively transitioning away from fossil fuels toward an electrified transportation future [3]. However, the success of this transition is heavily dependent on the development of robust EV charging infrastructure. Proactive and strategic deployment of charging stations is indispensable for enabling a smooth shift to electric mobility [4].
Despite recent progress, including a notable 55% increase in public charging stations reported by the International Energy Agency’s Global EV Outlook 2023 [5], significant challenges persist. These include determining optimal locations for charging stations, managing service coverage and efficiency, and ensuring the economic viability of charging infrastructure investments [6]. According to [7], residents’ satisfaction with the availability of charging stations significantly influences their willingness to adopt EVs. Leading EV manufacturers have advocated for increased government incentives and substantial investments in charging infrastructure to boost EV adoption rates worldwide [8]. Conversely, recent reports [9,10,11] suggest that the rapid expansion of public EV charging infrastructure across different countries has led to overcapacity, resulting in inefficiencies and underutilized charging stations. These misalignments with actual EV user demand have caused resource waste and run counter to sustainability. Therefore, thorough planning and deployment of EV charging infrastructure are crucial for providing quality charging services and promoting a sustainable and efficient charging ecosystem. Such an ecosystem not only supports the growing EV market but also ensures that the environmental and social benefits of EVs are fully realized, ultimately contributing to global efforts to sustainable development.
Numerous optimization models have been proposed to address these issues. Early work applied flow-based siting to locate facilities at centroids of traffic demand [12], while operational studies optimized EV assignment and charging for fixed stations using equilibrium formulations [13]. More recent efforts attempted to link investment and operations, such as facility-level joint designs [14], joint optimization of autonomous EV fleets with station siting [15], and bi-level formulations integrating location and routing [16]. These advances (see Section 2 for a full review) nevertheless fall short in two critical respects. First, most inadequately capture the spatialtemporal coupling of charging, even though real-world charging demand varies strongly across both time and space. Second, building such detailed models typically requires extensive expert effort to define problem structures, extract relevant features, and handcraft optimization formulations. These limitations hinder efficient and scalable planning at city scale.
To overcome this modeling bottleneck, recent advances in large language models (LLMs) provide promising tools to accelerate formulation. Several studies [17,18,19] have demonstrated that LLMs are capable of translating structured naturallanguage descriptions into optimization variables, constraints, and solver-ready code. Building on this capability, we propose a human-LLM collaborative workflow for EV charging infras-tructure planning. This approach reduces the burden of manual formulation while allowing experts to iteratively refine and validate model structures, enabling rapid development of contextspecific optimization models. However, even with faster model generation, realistic planning problems that account for finegrained spatial and temporal demand remain very large-scale and high-dimensional, creating significant computational challenges.
Addressing this computational complexity requires solution methods that scale beyond centralized solvers. Distributed optimization, particularly the Alternating Direction Method of Multipliers (ADMM) [20], is well suited to such settings. ADMM decomposes large models into parallelizable subproblems, reducing computational burdens while maintaining theoretical guarantees. It has already proven effective in diverse energy system applications, including microgrid energy trading [21], restoration in transmission systems with high renewable penetration [22], energy management for virtual power plants [23], and airconditioning system optimization [24]. By extending these advantages to EV infrastructure planning, we enable efficient solution of city-scale problems where spa
This content is AI-processed based on open access ArXiv data.