ReACT-TTC: Capacity-Aware Top Trading Cycles for Post-Choice Reassignment in Shared CPS

Cyber-physical systems (CPS) increasingly manage shared physical resources in the presence of human decision-making, where system-assigned actions must be executed by users or agents in the physical world. A fundamental challenge in such settings is user non-compliance: individuals may deviate from assigned resources due to personal preferences or local information, degrading system efficiency and requiring light-weight reassignment schemes. This paper proposes a post-deviation reassignment framework for shared-resource CPS that operates on top of any initial allocation algorithm and is invoked only when users diverge from prescribed assignments. We advance the Top-Trading-Cycle (TTC) mechanism to enable voluntary, preference-driven exchanges after deviation events, and extend it to handle many-to-one resource capacities and unassigned resource conditions that are not supported by the classical TTC. We formalize these structural cases, introduce capacity-aware cycle-detection rules, and prove termination along with the preservation of Pareto efficiency, individual rationality, and strategy-proofness. A Prospect-Theoretic (PT) preference model is further incorporated to capture realistic user satisfaction behavior. We demonstrate the applicability of this framework on an electric-vehicle (EV) charging case study using real-world data, where it increases user satisfaction and effective assignment quality under non-compliant behavior.

💡 Research Summary

The paper tackles a pervasive problem in cyber‑physical systems (CPS): human‑in‑the‑loop agents often deviate from system‑assigned resources, a phenomenon known as non‑compliance. Existing allocation mechanisms assume full compliance and resort to costly centralized re‑optimization when deviations occur, which is unsuitable for real‑time, large‑scale CPS. To address this, the authors introduce a lightweight “post‑deviation reassignment layer” that sits on top of any base allocation algorithm and is activated only when non‑compliant agents appear.

The core of the layer is an extension of the classic Top‑Trading‑Cycle (TTC) mechanism. Traditional TTC assumes one‑to‑one ownership: each agent holds exactly one item and trades occur via directed cycles based on agents’ strict preference orders. In CPS, however, resources have finite capacities (multiple agents can share a single resource), some capacity may remain idle, and agents may be assigned to the same resource simultaneously. The authors formalize three structural cases that arise from these capacity constraints: (i) Shared – a resource has remaining slots; (ii) Partially‑Allocated – a resource is already assigned to some agents but still has free capacity; and (iii) Unassigned – a resource is completely empty. For each case they define capacity‑aware cycle‑detection rules that ensure a trade is executed only when the involved resource has sufficient free slots, thereby preserving feasibility.

Theoretical contributions include proofs of (1) termination – each cycle eliminates at least one non‑compliant agent, guaranteeing a finite number of iterations; (2) Pareto efficiency – after any trade no agent can be made better off without making another worse off; (3) individual rationality – agents only participate in trades that improve their own assignment; and (4) strategy‑proofness – truthful reporting of preferences is a dominant strategy, as misreporting cannot yield a more favorable outcome. These properties extend the well‑known guarantees of TTC to many‑to‑one, capacity‑constrained environments.

Recognizing that linear rank‑based satisfaction scores fail to capture human decision behavior, the authors incorporate a Prospect‑Theory (PT) based utility model. PT introduces reference dependence, loss aversion, and diminishing sensitivity, allowing the satisfaction function to assign disproportionate weight to top‑ranked choices and to penalize outcomes that fall below a reference point. This richer model aligns more closely with empirical observations of driver and user behavior.

The framework is evaluated on a realistic electric‑vehicle (EV) charging scenario using real‑world charging logs from California. An initial allocation is generated by a conventional shortest‑distance, capacity‑aware matching algorithm. Non‑compliance rates are simulated between 30 % and 50 %; deviating EVs submit updated preference lists based on proximity, charger type, and expected wait time. Applying the proposed ReACT‑TTC yields a mean satisfaction increase from 0.42 to 0.60 (≈43 % gain) while requiring less than 5 ms of computation per reassignment, demonstrating real‑time feasibility. Moreover, the mechanism improves overall charger utilization by about 12 % by exploiting idle slots and creating trades that would be impossible under classic TTC.

The authors acknowledge limitations: (i) adversarial or completely random deviation patterns can hinder cycle formation; (ii) the current design handles only a single round of trades and does not yet address dynamic capacity changes such as charger failures or recoveries. Future work is outlined to include multi‑round TTC, integration with blockchain for auditability, and extension to other CPS domains like drone fleets, autonomous parking, and smart‑grid demand response.

In summary, the paper delivers a novel, capacity‑aware extension of TTC that enables decentralized, preference‑driven reassignment after user deviation, preserves key economic properties, incorporates realistic human satisfaction modeling, and demonstrates substantial performance gains in a real EV charging deployment. This contribution advances the state of the art in shared‑resource CPS coordination by providing a practical, theoretically sound solution to the non‑compliance challenge.