Online Decision-Making Under Uncertainty for Vehicle-to-Building Systems

Rishav Sen, Yunuo Zhang, Fangqi Liu, Jose Paolo Talusan, Ava Pettet, Yoshinori Suzue, Ayan Mukhopadhyay, Abhishek Dubey

Proceedings of the ACM/IEEE 16th International Conference on Cyber-Physical Systems (ICCPS) 2025 ICCPS '25

PDF

Why This Matters

Vehicle-to-building systems present a unique control challenge where centralized optimization traditionally assumes full system knowledge, yet real deployment requires online decision-making under uncertainty in EV arrivals, pricing, and building loads. This work is innovative because it explicitly models the online decision-making nature of V2B coordination, demonstrating how MCTS-based planning can provide near-optimal decisions in high-dimensional, uncertain environments without requiring expensive offline computation.

What We Did

This work formulates and solves the vehicle-to-building charging problem as a Markov decision process, emphasizing the challenge of real-time decision-making under uncertainty. The research models the problem using online MCTS-based approaches to handle dynamic electricity pricing, heterogeneous EV chargers, and stochastic EV arrivals. The work integrates domain-knowledge guided exploration with Monte Carlo tree search to enable efficient decision-making that balances immediate operational constraints with long-term energy cost minimization.

Key Results

Evaluation using real EV data and building information demonstrates that the online MCTS approach achieves significant improvements in total electricity costs while meeting all user charging requirements. The framework shows that principled online decision-making substantially outperforms greedy heuristics and provides better real-world applicability than offline optimization.

Full Abstract

Cite This Paper

@inproceedings{sen2025iccps,
  author = {Sen, Rishav and Zhang, Yunuo and Liu, Fangqi and Talusan, Jose Paolo and Pettet, Ava and Suzue, Yoshinori and Mukhopadhyay, Ayan and Dubey, Abhishek},
  booktitle = {Proceedings of the ACM/IEEE 16th International Conference on Cyber-Physical Systems (ICCPS)},
  title = {Online Decision-Making Under Uncertainty for Vehicle-to-Building Systems},
  year = {2025},
  address = {New York, NY, USA},
  publisher = {Association for Computing Machinery},
  series = {ICCPS '25},
  abstract = {Vehicle-to-building (V2B) systems combine physical infrastructure such as smart buildings and electric vehicles (EVs) connected to chargers at the building, with digital control mechanisms to manage energy use. By utilizing EVs as flexible energy reservoirs, buildings can dynamically charge and discharge EVs to effectively manage energy usage, and reduce costs under time-variable pricing and demand charge policies. This setup leads to the V2B optimization problem, where buildings coordinate EV charging and discharging to minimize total electricity costs while meeting users' charging requirements. However, the V2B optimization problem is difficult due to: 1) fluctuating electricity pricing, which includes both energy charges ($/kWh) and demand charges ($/kW); 2) long planning horizons (usually over 30 days); 3) heterogeneous chargers with differing charging rates, controllability, and directionality (unidirectional or bidirectional); and 4) user-specific battery levels at departure to ensure user requirements are met. While existing approaches often model this setting as a single-shot combinatorial optimization problem, we highlight critical limitations in prior work and instead model the V2B optimization problem as a Markov decision process, i.e., a stochastic control process. Solving the resulting MDP is challenging due to the large state and action spaces. To address the challenges of the large state space, we leverage online search, and we counter the action space by using domain-specific heuristics to prune unpromising actions. We validate our approach in collaboration with an EV manufacturer and a smart building operator in California, United States, showing that the proposed framework significantly outperforms state-of-the-art methods.},
  acceptance = {28.4},
  category = {selective},
  contribution = {lead},
  location = {California, USA},
  numpages = {10},
  ranking = {rank1},
  keywords = {vehicle-to-building, EV charging, online optimization, Monte Carlo tree search, stochastic decision-making, demand charge management}
}
Quick Info
Year 2025
Series ICCPS '25
Keywords
vehicle-to-building EV charging online optimization Monte Carlo tree search stochastic decision-making demand charge management
Research Areas
energy CPS planning
Search Tags

Online, Decision, Making, Uncertainty, Vehicle, Building, Systems, vehicle-to-building, EV charging, online optimization, Monte Carlo tree search, stochastic decision-making, demand charge management, energy, CPS, planning, 2025, Sen, Zhang, Liu, Talusan, Pettet, Suzue, Mukhopadhyay, Dubey, ICCPS25