Chapter 11 - Integration

Abstract

The chapter introduces blackboard system as a development environment. This chapter explains how different concepts and techniques in artificial intelligence can be integrated to develop an intelligent system. One approach to integrating the component parts of intelligent systems involves the concept of blackboard systems. This approach is simple and sufficiently powerful to enable quite complex machines and systems to be developed. The central feature of the blackboard system is an area of working memory called the blackboard. The knowledge and data stored or written “on the blackboard” is intended to be public and accessible to any one of a set of independent agents. Any agent can write to the blackboard and read from the blackboard. This public information is not necessarily managed by any of the agents, and may emerge from their interaction. This architecture allows each of the agents to do its own business with its only external interface being the blackboard. This greatly simplifies the conceptual nature of each agent, and what one needs to know about its interactions with other agents. The chapter explains how to run many rule-based systems in parallel. Concepts related to implementing a perception subsystem are also explained in detail in the chapter.