Knowledge-level modeling based on TRSA method

The paper proposes a knowledge-level modeling method, called TRSA, which describes a task structure (i.e. the abstract structure for a problem-solving task) by means of task reduction and describes the reasoning activities for the subtasks included in the task structure by means of subtask association. As conceptual models of problem-solving processes, task structures can be constructed dynamically according to the features of application domains and the requirement of problem solving tasks. Hence, they can become the constraints and guide for acquiring reasoning knowledge and maintaining knowledge bases. In addition, a visual, interactive, structural modeling language and a symbol-level functional architecture are provided to support TRSA.<>