A Predictive Simulation Model of Competitive Dynamics in Innovation
Active from 2009 - 2013
This interdisciplinary project contributes to two scientific areas. First, in order to model environments with competitive dynamism, new computational search techniques will be developed. Prior research has focused on single-agent and team search, but not on situations where multiple competing agents search for the same set of peaks simultaneously. The proposed project aims to fill this gap: Mechanisms will be developed for making knowledge hidden, shared, and public, and for dealing with landscapes that change dynamically based on other agents' search. The end result will be a set of general algorithms for competitive multiagent search and a theory of when they are effective.

Second, the main insight for organizational theory is that innovation is a competitive activity. While prior research has concentrated on mechanisms that are internal to the firm, companies do not search in isolation: Instead, they distinguish themselves from their competitors, learn from each other, and their inventions crowd and boost each other. These interactions will be incorporated into an evidence-based simulation model, creating a competitive environment that represents complex real-world interactions more comprehensively than previous approaches. It also makes it possible to identify situations in which firms do not behave rationally (i.e. optimally) and helps explain why. The end result will be an experimentally verified computational theory of how companies innovate in a competitive environment.

This research is supported by the National Science Foundation under grant SBE-0914796, with Riitta Katila of Stanford University as a Co-PI.

Erkin Bahceci Ph.D. Student erkin [at] cs utexas edu
Risto Miikkulainen Faculty risto [at] cs utexas edu
Riitta Katila Collaborator rkatila [at] stanford edu
NKVis This package contains a 3D visualization tool for NK fitness landscapes. Two types of visualizations are provided: a 2011