Dimensions And Heuristics For The Design Of Distributed Artificial Intelligence Architectures
L.M. Lewis, R.C. Gwin · 1991
Our objective in the paper is to establish a framework for thinking about problems whose solutions require multiple agents working together. Particularly, we are concerned with agents who operate in an undersea environment. We discuss three problems within this framework: (i) semi-automated platform command and control, (ii) fully-automated control of autonomous underwater vehicles (AUVs), and (iii) joint platformlAUV mission control. Our work proceeds under the hypothesis that each problem is a Distributed Artificial Intelligence (DAI) problem and that each problem is essentially isomorphic. In the paper we characterize the concept of DAI problem, propose a table of dimensions for casting DAI problems, discuss several DAI design heuristics, and describe our work on the three problems above.