Agentic aggregator for electric bus fleets: pricing, trade-offs

Jônatas Augusto Manzolli, Ali Eslami, Luis Miranda-Moreno and Jiangbo Yu submitted an agentic aggregator framework for electric bus fleet operations to arXiv on 24 Jun 2026 (arXiv:2606.26400). The paper couples an optimization-based electric bus scheduling model with supervisory agents that detect disturbances, adapt tariffs and evaluate schedules to coordinate charging, battery state-of-charge, charger availability, electricity prices, route-energy uncertainty and vehicle-to-grid opportunities.

What is the agentic aggregator framework and how does it work?

The framework pairs an optimization core that enforces physical feasibility across routes, chargers, batteries and V2G exchanges with an agentic layer that interprets operating conditions, triggers real-time re-optimization and defines how flexibility value is shared between the aggregator and the public transport operator (PTO). The optimization core is responsible for feasible schedules across routes, chargers, batteries and V2G exchanges; the supervisory agents perform disturbance detection, tariff adaptation and schedule evaluation and can activate re-optimization when operating conditions change.

That design separates low-level physical constraints from higher-level decision logic: the optimizer keeps route and charging plans physically valid while agents monitor service delays, route-energy deviations and electricity price shocks and decide when to re-run or adjust plans. The agentic layer also determines allocation rules for flexibility value between aggregator and PTO, so pricing behavior is part of the coordination logic, not only a system output.

How did the framework perform in the depot case study?

In a realistic depot case study the authors evaluated day-ahead and real-time operations under two coordination modes, profit-based and operation-based, and considered service delays, route-energy deviations, electricity price shocks and combined disturbances. The experiments show that agentic aggregation maintained feasible schedules, activated re-optimization selectively and improved the use of charging and V2G flexibility while operating under both day-ahead and real-time horizons.

The paper finds a practical trade-off: when the aggregator is configured around profit-oriented pricing, the same agentic capability that reduces operational complexity can extract value from the PTO. In contrast, an operation-based coordination mode prioritizes service reliability and operational constraints over profit-driven tariff-setting. The case study therefore contrasts profit-based and operation-based modes across the same disturbance scenarios to reveal how pricing choices change who captures value from flexibility.

Why it matters

The framework addresses a concrete coordination challenge: electric bus fleets interact with variable electricity prices, charger constraints and uncertain route energy needs while offering V2G opportunities. By combining an optimization core with supervisory agents, the design lets systems react in real time without violating physical constraints. That can improve charging and V2G utilization and preserve schedule feasibility during price shocks or delays. At the same time the results highlight a governance issue: agentic aggregation can shift economic value depending on tariff and coordination settings, which matters for public-fleet operators and regulators.

The paper therefore links technical capability to policy choices, arguing that deployment in public-fleet contexts needs transparent coordination modes, auditable tariff-setting and explicit value-sharing rules to prevent unintended transfers of value from PTOs to profit-seeking aggregators.

What to watch

Look for follow-up work or deployments that publish concrete tariff rules, audited value-sharing mechanisms and comparisons of profit-based versus operation-based coordination in live depots. The next milestones that would confirm the paper's implications are demonstrations of the framework in operational depots and publicly disclosed tariff/allocation contracts that show how flexibility payments are split between aggregators and PTOs.

Details: the paper appears on arXiv as arXiv:2606.26400 and was submitted on 24 Jun 2026 by Manzolli, Eslami, Miranda-Moreno and Yu. The authors tested day-ahead and real-time operations and contrasted profit-based and operation-based coordination modes under service delays, route-energy deviations, electricity price shocks and combined disturbances.