Systems simulation. Supplementing global models with computational models: An assessment and an energy example

Many global modeling efforts have been criticized for not paying enough attention to political decision making. Computational models, based on much of the methodology developed by scholars in the “artificial intelligence” community have been utilized in recent years to represent some aspects of political decision making. Could intellectual progress be made by combining some of the features of these heretofore disparate academic strands? Answering that question is the first emphasis of this paper. Four criteria are suggested under which supplementing global modeling efforts with computational models seems to make sense. The criteria are the identity of the political unit being modeled, the depth of the modeler's understanding of the political unit's problem‐solving behavior, the question of whether the political unit can be reasonably characterized as exhibiting an identifiable general mode of problem solving, and whether there is sufficient information available to provide for a validity test. A mechanism‐elucidating model of the U.S. congressional energy research and development decision making is presented in the second portion of the paper. Programmed in LISP, the model is presented to be suggestive of the type of efforts which could fruitfully supplement global models. The intentional inferencing procedures through which a “modal” and two other significant representative members of the U.S. Congress decide how to vote on legislation in the area of funding energy research and development is the subject matter of the LISP‐based example. The model is tested for the case of 1985 congressional legislation concerning funding of the Synthetic Fuels Program, and is supported. Conclusions of the model are compared with those typically input to a global model, and the computational model is found to be more accurate because it suggests lower levels of U.S. energy research and development funding. The substitution of an emphasis on intentional inferencing of key political actors for the common emphasis of national economic rationality is the central change leading to this improvement. Levels dealt with in this article include organization, group, society, and supranational system. The decider is the major subsystem involved.