Robust Software with Errors through Team-Oriented Programming

University Of South Carolina At Columbia, Columbia SC

Investigators

Abstract

This CONACyT research project will enable a research group at the University of South Carolina to cooperate with a research group at the Instituto Politecnico Nacional in Mexico City, Mexico. Together they will develop a methodology that enables software agents with different ontologies to interact and exchange information. The objectives are to increase the ease with which such software can be produced and its robustness. Under this approach, software development would consist of choosing among "volunteers" (agents from an active repository) and assembling them into a problem-solving team, instead of debugging lines of code. Software development would be more like coaching and less like traditional programming. The research will explore 1) how independently constructed autonomous agents reach mutual understanding, 2) what categories of software are suitable for team-oriented programming, and 3) how team behaviors can be analyzed and validated. Groups of students will assemble agent-based software systems and compare the resulting systems for robustness with software developed conventionally. The two resarch sites will collaborate through online meetings, the exchange of graduate students and mutual visits to each other's research centers.