Network Games: An Approach to Reasoning in An Adversarial Environment
Abstract
A basic empirical fact about foreign policy and national security is that the decision makers within political and military organizations (henceforth called entities) are constantly faced with Ònetworked environmentsÓ with both cooperative and adversarial elements. By a networked environment is meant an environment whose entities are connected together through a web of cooperative and uncooperative relations. Using mainly concepts from game theory, this project aims to develop tools to help predict what types of decisions these decision makers might be led to make when confronted with such an environment. The proposed project consists of mainly three thrusts. The first thrust is to examine how decision makers allocate power within a static or dynamic networked environment possessing adversarial and cooperative entities. Toward this end, three power allocation games for three different scenarios will be developed, with the first being a static game where agents protect the survival of their friends and oppose that of their adversaries in everyday politics, the second being a dynamic game where decision makers iteratively update their power allocation strategies to maximize spoils in multiple battlefields or theaters in combat, and the third being a differential game where decision makers optimally control their power allocation paths in a dynamic environment with a grand, long-term strategy in mind. The second thrust is to develop tools to make predictions based on the outcomes of the games considered in the preceding thrust. Certain historic cases such as World Wars I and II will be explored in an effort to validate the predictive tools under consideration. For example, could the outcome of World War II have been predicted using these tools? These tools of course might also be used to make predictions relevant to contemporary events which have not yet taken place. For example, as a demonstration of the tools potential utility, an effort will be made to develop predictions of the possible outcomes of various military, political and economic actions which might be taken by the United States and its allies toward North Korea. The third thrust is to study how decision makers might act to control the networked environments within which they operate to obtain better power allocation outcomes. This will be approached using three models, namely a static game, a dynamical system and a differential game. The three models will then be used in an effort to solve unresolved puzzles in political and military affairs, such as the causes of famous wars in human history and great powers optimal design of the world system. The central aim of this project is to construct and extend an appropriate mathematical framework for modeling decision makers strategic reasoning and behaviors in networked, adversarial environments and for demonstrating these models predictive capabilities. This will be achieved by combining the concepts and methods from game theory, the social sciences and the system sciences. The project will serve a full range of causal, explanatory and predictive purposes. New technical problems will be formulated, whose solutions have the potential of contributing to the mechanisms of decision-making in a culturally and ideologically diverse environment.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 14, 2019
- Source ID
- W911NF1910035
Entities
People
- A. Stephen Morse
Organizations
- Army Contracting Command
- United States Army
- Yale University