General Economic Equilibrium as a Unifying Concept in Energy-Economic Modeling
比较了四种主流能源经济模型（变系数投入产出、过程网络、线性规划、非线性优化）相对于一般经济均衡模型的简化假设，帮助理解通用性与计算可解释性之间的权衡。
In the pristine model of general economic equilibrium producers and consumers are assumed to take prices for their inputs and outputs as given. A market equilibrium solution is obtained when the prices of all products lead to equal amounts of supply and demand for each. The mathematical properties of this model provide a rich and powerful unifying foundation for much of modern microeconomic theory. However, in many practical applications it is difficult or impossible to develop the data, and formulate and solve the equations required to implement the model of general economic equilibrium in its most general form. In practice, generality of formulation is often sacrificed for ease of computation and interpretation. These tradeoffs allow the latest breakthroughs in optimization algorithms and computer technology to be applied in the analysis of important societal problems. A potential drawback associated with this otherwise desirable trend is that the analysis could become infatuated with the use of particular algorithms, and lose sight of the restrictive simplifying assumptions they imply. The present paper includes a comparison of the simplifying assumptions required in four of the most popular types of energy-economic models with respect to the model of general economic equilibrium. This comparison helps sharpen our appreciation for the tradeoffs between generalization of formulation and ease of computation and interpretation that are available. The concept of general economic equilibrium is employed to provide a common framework for four ostensibly different approaches to large-scale modeling: (1) variable-coefficient input-output theory, (2) process network methodology, (3) linear programming and (4) general nonlinear optimization. The similarities and the differences of the four approaches are isolated within this framework. This comparison makes both the absolute and the relative strengths and weaknesses of the models more transparent.