异质发电商竞争下电力市场减排锦标博弈--结构与行为

Mitigation tournament game under heterogeneous generators competition in electricity market: Structure and behavior

利用行为经济学理论研究了电力市场中异质发电商竞争下的减排锦标博弈问题,先后讨论了不同市场与奖金结构下均衡的边际获胜概率及其最优减排努力程度.从实验经济学角度对理论模型进行了实验检验,在此基础上将模型拓展到考虑影响发电商减排决策的非金钱因素,如社会比较的减排锦标模型,并给出了模型的最优参数估计与均衡预测值.通过比较分析发现,在3、4人减排锦标赛中,针对不同发电主体构成,增加胜利者奖励数量既不能驱动强者提高也不会迫使弱者降低各自的最优减排量;无论初始调度禀赋如何,只要检验实验信息完全公开,所有发电商均存在过度减排投资行为,且在双强者单弱者参与情形下,强势发电商的减排努力水平与获胜者数量正相关关系与标准理论预测相悖;对参数的约束条件显著降低行为经济学模型拟合度,其中广义模型均衡预测最契合检验实验的基本特征,而特定嵌套模型验证理论预测的可行性.

The mitigation tournament game under heterogeneous generators competition in electricity markets is studied by using behavioral economics. The equilibrium marginal probabilities of winning and the optimal mitigation efforts in different market and bonus structures are discussed successively. Theoretical models are experimentally tested from the perspective of experimental economics and the models are extended to tournament models of mitigation considering non-pecuniary factors such as social comparison which also affects a generator＇ s mitigation decision. The optimal parameter estimations and equilibrium predictive values of generalized models are given. The results gained through comparative analysis indicate that increasing the number of winning prizes, regardless of the generation subject compositions, neither drives superiors to raise nor forces the vulnerable to decrease respective mitigation quantities in a tournament with three or four participants. No matter what the initial dispatching endowments are, all generators present overinvestment behaviors as long as the experiment information is completely public. Contrary to the standard theory predictions, the levels of the latter＇ s mitigation efforts are positively correlated with number of winners when a single vulnerable and double superior generators compete. The parameter constraints reduce the fitness of the behavioral economics model significantly and the equilibrium prediction of generalized model fits the basic features of test experiment best in contrast with the specific nested model which verifies the feasibility of the theoretical predictions.