We propose a refinement process of dynamic equilibria based on small random perturbations (SRPs) of the backward perfect foresight (bpf) equilibrium map in a class of one-step, forward-looking dynamic models. An equilibrium is selected if its stationary measure is the limit of the stationary measures associated with the processes generated by the SRPs of the bpf maps, as the perturbation size approaches 0. We show that, for full measure sets of parameter values of a large class of one-parameter families of unimodal bpf maps, only determinate cycles or the chaotic sunspot equilibrium defined by Araujo and Maldonado (2000) is selected. Two examples are provided illustrating such refinement process.

中文翻译：

---

使用小的随机扰动优化动态平衡

我们提出了一种基于后向完美预见 (bpf) 平衡图的小随机扰动 (SRP) 的动态平衡的细化过程，该过程是一类单步、前瞻性动态模型。当扰动大小接近 0 时，如果其平稳测量是与 bpf 映射的 SRP 生成的过程相关的平稳测量的极限，则选择平衡。我们表明，对于大类的参数值的完整测量集在单峰 bpf 映射的单参数族中，仅选择确定周期或由 Araujo 和 Maldonado (2000) 定义的混沌太阳黑子平衡。提供了两个示例来说明这种细化过程。