作者: Keith Decker , Victor Lesser
DOI:
关键词: Task analysis 、 Machine learning 、 Task (project management) 、 State (computer science) 、 Computer science 、 Context (computing) 、 Structure (mathematical logic) 、 Cooperative distributed problem solving 、 Control (management) 、 Artificial intelligence 、 Domain (software engineering)
摘要: Formal approaches to specifying how the mental state of an agent entails that it perform particular actions put at center analysis. For some questions and purposes, is more realistic convenient for analysis be task environment, domain, or society which agents will a part. This paper presents such environment-oriented modeling framework can work hand-in-hand with agent-centered approaches. Our approach features careful attention quantitative computational interrelationships between tasks, what information available (and when) update agent's state, general structure environment rather than single-instance examples. A model used both simulation; avoids methodological problems relying solely on examples, provides concrete, meaningful characterizations theories. give example in context cooperative distributed problem solving, but our analyzing centralized parallel control as well.