Abstract
This paper presents An Intelligent Agent Model for a Given Task in a Specified Environment. The methodology in this work is based on mixing algorithmic and function approaches to construct the intelligent agent model. The paper concentrates on building an intelligent agent model as a knowledge-based system interacting with dynamic environment to perform tasks.
The class structure used to represent the environment in the knowledge base depends on three types of knowledge representation forms: production rule, semantic net, and frames. Each object in the environment is an instance of the class environment. Algorithms and functions are used for gaining knowledge from the state space of environment to build the task. The intelligent agent model can understand the environment from any position and can detect many subtasks, arrange them in a queue for execution, and has the ability to make a decision at a high level of thinking.
The intelligent agent model is able to calculate the persistent changes in the external dynamic environment and any sudden change, such as observing the existence of any obstacle in the environment and avoiding it. The intelligent agent is also able to learn and take the reasonable decision in the dynamic environment and automatically select action based on task characteristics. Therefore, the intelligent agent can solve many different types of problems.