I-Corps: Software Infrastructure for Cloud Simulation and Strategic Analysis

Regents Of The University Of Michigan - Ann Arbor, Ann Arbor MI

Investigators

Abstract

The availability of large-scale computing provided by platforms reduces the costs of exploiting compute-intensive analysis in support of strategic decision making. Computational infrastructure developed by the project team provides middleware that shields the user from such concerns, and thus can provide a valuable gateway to the cloud for simulation-based decision support systems. The approach is founded on an emerging methodology for strategic reasoning based on empirical game-theoretic analysis (EGTA). EGTA combines massive simulation and other empirical methods with concepts from game theory to support decisions that anticipate strategies and reactions from other agents (e.g., competitors, customers) in a complex environment. The approach aligns the simplicity of agent-based modeling with the principled analytic basis of game theory. The technology developed promises to dramatically increase the scale and fidelity of EGTA studies, extending the scope of game-theoretic reasoning from stylized domains to real-world problems. Many problem environments faced by firms, individuals, and society are fundamentally strategic, in that the outcome of one agent's action depends pivotally on the choices of others. Improving strategic decision making in complex domains offers benefits for many economically important applications, for example, supply chain risk management, infrastructure and location planning, and technology adoption and standardization. Facilities developed in this project will broaden access to large-scale distributed computing resources, leading to significant efficiencies in utilization of computational infrastructure.