随着代理人工智能系统越来越多地应用于网络物理环境,，它们的评估需要评估任务绩效和可信度。在去中心化能源市场,，自主代理可以提高市场效用,，但也可能利用无效的物理数据,创造人为流动性,并产生不稳定的治理决策。因此,我们提出SolarChain-Eval,作为评估值得信赖的经济主体的物理约束基准。它将市场治理制定为与 Gymnasium 兼容的马尔可夫决策过程,，其中代理每小时做出决策。 SolarChain-Eval 从多个维度, 评估每项政策，包括市场效用, 物理安全, 滑点, 行动平滑性, 空间公平, 和可审计性。为了支持代理评估, SolarChain-Eval 结合了基于 LLM 的 Planner/Auditor 层。计划员定义事件级别的操作界限和审核规则,，而审核员则审查和修改高风险操作。所有干预措施均通过结构化日志, 记录，包括触发信号, 建议的行动, 修订的行动, 和审计理由。静态, 随机, 近视, RL, 和RL+LLM 策略的实验揭示了明显的效用与安全权衡。 RL 代理提高了市场效用，但仍然会产生不安全行为。当物理惩罚被消除时，, 奖励最大化代理会利用无效生成并增加人工流动性。 LLM Planner/Auditor 提高了可审计性并减轻了选定的风险,，但它无法完全补偿错误指定的奖励函数。这些结果表明，值得信赖的代理人工智能评估需要物理约束和透明的干预痕迹。我们在 GitHub 上以开放获取方式发布数据和代码，以实现可复制性。
As agentic AI systems are increasingly applied to cyber-physical environments, their evaluation requires assessment of both task performance and trustworthiness. In decentralized energy markets, autonomous agents may improve market utility, but may also exploit invalid physical data, create artificial liquidity, and produce unstable governance decisions. Therefore, we propose SolarChain-Eval, a physics-constrained benchmark for evaluating trustworthy economic agents. It formulates market governance as a Gymnasium-compatible Markov Decision Process, where agents make hourly decisions. SolarChain-Eval evaluates each policy across multiple dimensions, including market utility, physical safety, slippage, action smoothness, spatial fairness, and auditability. To support agentic evaluation, SolarChain-Eval incorporates an LLM-based Planner/Auditor layer. The Planner defines episode-level action bounds and audit rules, while the Auditor reviews and revises high-risk actions. All interventions are recorded through structured logs, including trigger signals, proposed actions, revised actions, and audit rationales. Experiments with static, random, myopic, RL, and RL+LLM policies reveal a clear utility-safety trade-off. RL agents improve market utility but can still produce unsafe behavior. When the physics penalty is removed, reward-maximizing agents exploit invalid generation and increase artificial liquidity. The LLM Planner/Auditor improves auditability and mitigates selected risks, but it cannot fully compensate for a misspecified reward function. These results indicate that trustworthy agentic AI evaluation requires both physical constraints and transparent intervention traces. We release data and code as open access on GitHub for replicability.
科目: 人工智能 (cs.AI); 新兴技术 (cs.ET); 机器学习 (cs.LG); 多代理系统 (cs.MA); 一般经济学 (econ.GN)
Subjects: Artificial Intelligence (cs.AI); Emerging Technologies (cs.ET); Machine Learning (cs.LG); Multiagent Systems (cs.MA); General Economics (econ.GN)