When we aggregate distributed, flexible resources—such as a collection of electric vehicles, smart home appliances, or a shared fleet of delivery trucks—we often expect the resulting system to act as a highly efficient, macro-scale buffer. The intuitive hope is that pooling these assets will smooth out individual volatility, allowing an operator to maximize both the capacity of the system and its real-time responsiveness.
However, a deeper look at the physics and mathematics of aggregation reveals a fundamental barrier: a strict, non-linear trade-off between resource utilization (or capacity) and operational flexibility (or response rates). Pushing an aggregated system to its limit in one dimension inherently collapses its capability in another.
The State-Space Conflict in Virtual Batteries
The most mathematically explicit version of this trade-off appears when trying to make a collection of flexible, deferrable energy loads mimic a standard chemical battery. In Battery Capacity of Deferrable Energy Demand, Daria Madjidian et al. prove that aggregated flexible loads cannot perfectly replicate a physical battery due to a fundamental conflict in their internal states.
To understand why, consider what is required to maximize a virtual battery's instantaneous performance:
- To maximize charging rate, the underlying devices must be at a low energy level (empty), leaving maximum headroom to absorb power from the grid.
- To maximize discharging rate (or consumption reduction), the devices must be at a high energy level (full), leaving maximum room to curtail their consumption without violating user constraints.
Because individual assets cannot be simultaneously full and empty, the aggregate system faces a physical boundary. As Daria Madjidian et al. demonstrate in Emulating batteries with deferrable energy demand, this limitation is governed by "slack"—the temporal flexibility available to shift loads.
When you attempt to maximize the total energy volume stored in this virtual battery, you consume this slack entirely. In Energy Storage From Aggregate Deferrable Demand: Fundamental Trade-Offs and Scheduling Policies, Daria Madjidian and Mardavij Roozbehani formalize this boundary condition: operating an emulated battery at its absolute maximum energy volume forces either its maximum charge rate or its maximum discharge rate to drop to zero. The feasible operating region is not a simple box, but a curved space where capacity and rate limits are deeply coupled.
Maximum Charge Rate
^
| * (High charge rate requires empty state)
| \
| \ <- Feasible Operating Boundary
| \
| * (High capacity forces rates to zero)
| \
+--------*--------> Total Energy Capacity / Volume
The Slack Trade-Off in Fleet Management
This tension between capacity utilization and flexibility is not unique to power grids; it governs logistics and mobility networks as well. In fleet management, maximizing vehicle utilization means minimizing idle time, which systematically strips away the operational slack required to handle unexpected demand or secondary tasks.
In long-haul trucking, Philip Kilby and Tommaso Urli explore this dynamic in Fleet Design Optimisation from Historical Data Using Constraint Programming and Large Neighbourhood Search. The authors show that while downsizing a fleet significantly boosts vehicle utilization during periods of lower demand, it leaves the system with zero buffer during peak periods, forcing a stark trade-off between asset efficiency and service reliability.
A similar mechanism occurs in ride-hailing platforms. In Book-ahead & supply management for ridesourcing platforms, Cesar N. Yahia et al. analyze the balance between advance reservations and spontaneous, on-demand requests. Locking in drivers for advance bookings allows operators to optimize schedules and minimize the total fleet size. However, this commitment locks up the system's state, removing the real-time flexibility needed to absorb sudden spikes in on-demand passenger requests without violating service guarantees.
Where Transport and Grid Flexibility Collide
The ultimate manifestation of this trade-off occurs when a single physical asset is asked to navigate both domains simultaneously. Vehicle-to-Grid (V2G) car-sharing networks present a scenario where vehicles must serve as both transport assets and grid-stabilizing virtual batteries.
In Vehicle-to-grid for car sharing - A simulation study for 2030, Nina Wiedemann et al. simulate this interaction at a national scale. They uncover a direct conflict between transport utilization and energy flexibility. When a car-sharing fleet is highly optimized for passenger trips, vehicles are constantly in motion or distributed across the network, leaving few idle cars plugged into charging stations.
Without idle vehicles sitting at chargers, the fleet loses its "slack" and cannot provide reliable capacity to the grid. To unlock meaningful grid stabilization services, the authors demonstrate that operators must deliberately maintain a larger fleet than transport demand requires. To gain power flexibility, they must intentionally accept lower vehicle utilization.
Key Foundational Work
Energy Storage From Aggregate Deferrable Demand: Fundamental Trade-Offs and Scheduling Policies (2017) — contributed 4 concept elements to this direction.
Battery Capacity of Deferrable Energy Demand (2016) — contributed 4 concept elements to this direction.
Emulating batteries with deferrable energy demand (2017) — contributed 3 concept elements to this direction.