Solving Non-Linear Models by Parametrizing Expectations

本项目开发了一种经济动态模型的求解技术。这种类型的求解程序使得在计算机上模拟模型成为可能。动态模型是随时间演变的模型，因此，它们最适合于经济政策分析。在计算机上模拟一个动态模型，可以分析模型的性质，并检验它是否能很好地描述实体经济。最终，计算机模拟可以确定经济政策中特定变化的影响，并可以为决策提供指导。在这个项目中，开发了一种寻找数值解的技术，并将其应用于几个模型。近年来，动态经济模型的数值求解已成为一个重要的问题。这些程序面临的主要挑战是计算的可跟踪性和准确性。计算可追溯性意味着算法允许快速计算，因此无需在计算机上花费太多时间和金钱就可以模拟有趣的模型。此外，由于所有这些方法都只是近似真实的解，因此有必要确保以合理的精度计算解。对于一类重要的模型，本项目开发的方法可能会在计算速度和精度方面改进现有方法。

This project develops a solution technique for economic dynamic models. A solution procedure of this type makes it possible to simulate models on a computer. Dynamic models are ones that evolve through time and, thus, they are most appropriate for economic policy analysis. Simulating a dynamic model on the computer makes it possible to analyze the properties of the model, and to test if it is a good description of the real economy. Ultimately, computer simulations can determine the effect of a particular change in economic policy, and they can provide guidance in policy-making. In this project a technique for finding numerical solutions is developed and applied to several models. Recently, finding numerical solution procedures for dynamic economic models has become an important issue. The main challenges to these procedures are computational tractability and accuracy. Computational tractability means that the algorithms allow for fast computation, so that it is feasible to simulate interesting models without having to spend too much time and money on the computer. Also, since all of these methods only approximate the true solution, it is necessary to make sure that the solution is computed with a reasonable degree of accuracy. For a class of important models, the method developed in this project is likely to improve on the existing methods both in terms of computational speed and in terms of accuracy.