Creation of DEVS models using imitation learning

Authors:
Michael W. Floyd;Gabriel A. Wainer
Affiliations:
Carleton University, Ottawa, ON, Canada;Carleton University, Ottawa, ON, Canada
Venue:
Proceedings of the 2010 Summer Computer Simulation Conference
Year:
2010

Citing 10
Cited 0

Case-based reasoning: foundational issues, methodological variations, and system approaches

AI Communications
Theory of Modeling and Simulation

Theory of Modeling and Simulation
CD++: a toolkit to develop DEVS models

Software—Practice & Experience
Robot Learning From Demonstration

ICML '97 Proceedings of the Fourteenth International Conference on Machine Learning
Learning for control from multiple demonstrations

Proceedings of the 25th international conference on Machine learning
ECD++ a DEVS based real-time simulator for embedded systems

Proceedings of the 2008 Spring simulation multiconference
Case-Based Planning and Execution for Real-Time Strategy Games

ICCBR '07 Proceedings of the 7th international conference on Case-Based Reasoning: Case-Based Reasoning Research and Development
Considerations for Real-Time Spatially-Aware Case-Based Reasoning: A Case Study in Robotic Soccer Imitation

ECCBR '08 Proceedings of the 9th European conference on Advances in Case-Based Reasoning
An Active Approach to Automatic Case Generation

ICCBR '09 Proceedings of the 8th International Conference on Case-Based Reasoning: Case-Based Reasoning Research and Development
Case-based plan recognition in computer games

ICCBR'03 Proceedings of the 5th international conference on Case-based reasoning: Research and Development

Quantified Score

Hi-index	0.00

Visualization

Abstract

Modelling and simulation of robotic control systems allows for low cost analysis and experimentation. However, creating these models requires a level of technical expertise. To improve the technical quality of such models, we propose a case-based reasoning approach to learn the behaviour of models using the DEVS formalism. By observing a desired behaviour, in the form of outputs produced in response to inputs, a DEVS model of the behaviour can be built. This model can then be used during simulation to imitate the behaviour of interest. Our results show that this learning approach can be used to successfully imitate the behaviour of several DEVS models. Additionally, it can be used to transfer behaviour to and from a non-DEVS model.