Any science requires the following elements: phenomena; empirical data that describe those phenomena; hypotheses at various levels to connect, explain and then predict those phenomena; and the ability to test those hypotheses, Once a hypothesis is able to predict it becomes a theory.
The science of innovation policy must be based on a platform that permits the formulation and testing of key hypotheses. Constructing such a platform is typically a challenge in the social sciences as there is often no ability to carry out experiments in the same way that the physical sciences can, i.e, in a laboratory. This project develops an approach to test hypotheses and theories by creating a computational laboratory within with hypothesis evolution and hypothesis testing can take place.
Intellectual Merit Much of the current research in the science of innovation policy is focused on collecting empirical data. There is a surprising paucity of testable models and theories. There are two directions from which models can come: empirical observation or inference based on a qualitative understanding of a field. Schumpeter, in his analysis of the success of capitalism, has produced one of the most explanatory qualitative theories of innovation known as ?creative destruction?. This theory of innovation takes a Lamarckian view of the evolution of products and processes, but it has yet to be formally modeled and tested. The contribution of this research is to advance the modeling and testing of Schumpeterian creative destruction.
In doing so the project creates a laboratory, which can then be used more generally. The laboratory is designed to synthesize concepts from the disparate fields of: innovation theory; creative production; computational sociology; social multi-agent systems; situated cognition; emergence; and data mining. Both the laboratory and the results it produces provide the foundations for a science of innovation policy: science that produces testable results, and one that can test hypotheses.
The laboratory uses computational sociology, a technology based on social multi-agent systems that allow for emergent behavior, as the modeling tool. Agents are generators and receivers of ?products? and take up novel, useful and unexpected products. The overall system behavior is structured to be emergent and is captured using data mining techniques. Because the system connects inputs to outputs at the overall system level, the effects of different types of innovation policies can then be tested in the laboratory.
Broader Impact The broader impacts of this research lie in multiple dimensions. The project will involve PhD students and give other students experience with this kind of integrative research. It will make connections to computer science, cognitive science, social science and design science. The results from this project provide as feedback to design and innovation educators initially at George Mason University and then to design and innovation educators at other universities through the use of demonstrations. The results from this project are disseminated via conference papers, journal papers and a website. The laboratory is publicly available publicly so that others can experiment with it through the website.
