Guiding Object Oriented Design Via the Knowledge Level Architecture: The Irrigated Wheat Testbed1

In this paper, we detail our current work on a crop management system for irrigated wheat in Egypt. The goal of our work is to develop a system that will address all aspects of crop management including varietal selection, planting/harvest date selection, sowing parameters decisions, insect/disease/weed identification and remediation, and irrigation/fertilization management. The approach we take to solve this problem is the Generic Task Approach to expert systems development pioneered by Chandrasekaran et al. By using the Generic Task (GT) approach, we set out to model the behavior of an expert in wheat crop management. To accomplish this goal, the GT approach builds on top of the object-oriented methodology and acts as a guiding overlay for analyzing knowledge intensive problems, such as wheat crop management. As a multitask problem, wheat crop management provides a testbed for the ideas of a Knowledge Level Architecture introduced by Sticklen. The Knowledge Level Architecture (KLA) provides a means of understanding large systems in terms of cooperating sub-agents. This paper describes the GT and KLA methodologies, focusing on the support they afford to the description and understanding of knowledge-based systems from an object-oriented perspective.