Artificial intelligence in general and software agents in particular are recognized as computer science disciplines that aim to model or simulate so-called intelligent human behaviors such as perception, decision-maki...Artificial intelligence in general and software agents in particular are recognized as computer science disciplines that aim to model or simulate so-called intelligent human behaviors such as perception, decision-making, understanding, learning, etc. This work presents an approach to designing a generic Intelligent Agent that can be used in a multi-agent system to solve a complex problem. The generic agent that is proposed can be instantiated as a concrete agent, which is enabled with learning and autonomy capabilities by using Artificial Neural Networks. To highlight the generic aspect, the proposition is instantiated to be used in agriculture, health and education. The instantiated software agent applied in agriculture can process images in real time and detect defect on plants’ leaf. In the health field, the agent process image to diagnose breast cancer. When applied in Education, the agent can load an image of a student’s script and grade it. The performance of the designed agent system has the same accuracy as that of the respective neural networks used to instantiate them. In the educational field, the software agent has an accuracy of 98.9% and in the health field, it has an accuracy of 99.56% while in the agricultural field, it has an accuracy of 97.2%.展开更多
This paper introduces the structure of a multiagent design system with machine learning mechanism and its application in mechanical design. Firs of are it introduces a hierarchical structure of the multiagent design ...This paper introduces the structure of a multiagent design system with machine learning mechanism and its application in mechanical design. Firs of are it introduces a hierarchical structure of the multiagent design system and takes a mechanical design system as an example. This structure provides a computational platform for cooperative design and sharing learning of multiple design agents. The paper analyses the principle of design activity and puts forward the architecture and learning mechanism of a design agent in datail. The architecture of a design agent is for providing support to learning activity and is based on the analysis of the design activity This is followed by a description of the design knowledge base framework and sharing learning process of multiagent. The main advantages of the system is that complex design task can be done by multiagent in a distributed environment and leaming results can be shared by a group of design agents. This system has partly been implemented in Visual C++ based on Mechanical Desktop 2.0 environment.展开更多
文摘Artificial intelligence in general and software agents in particular are recognized as computer science disciplines that aim to model or simulate so-called intelligent human behaviors such as perception, decision-making, understanding, learning, etc. This work presents an approach to designing a generic Intelligent Agent that can be used in a multi-agent system to solve a complex problem. The generic agent that is proposed can be instantiated as a concrete agent, which is enabled with learning and autonomy capabilities by using Artificial Neural Networks. To highlight the generic aspect, the proposition is instantiated to be used in agriculture, health and education. The instantiated software agent applied in agriculture can process images in real time and detect defect on plants’ leaf. In the health field, the agent process image to diagnose breast cancer. When applied in Education, the agent can load an image of a student’s script and grade it. The performance of the designed agent system has the same accuracy as that of the respective neural networks used to instantiate them. In the educational field, the software agent has an accuracy of 98.9% and in the health field, it has an accuracy of 99.56% while in the agricultural field, it has an accuracy of 97.2%.
文摘This paper introduces the structure of a multiagent design system with machine learning mechanism and its application in mechanical design. Firs of are it introduces a hierarchical structure of the multiagent design system and takes a mechanical design system as an example. This structure provides a computational platform for cooperative design and sharing learning of multiple design agents. The paper analyses the principle of design activity and puts forward the architecture and learning mechanism of a design agent in datail. The architecture of a design agent is for providing support to learning activity and is based on the analysis of the design activity This is followed by a description of the design knowledge base framework and sharing learning process of multiagent. The main advantages of the system is that complex design task can be done by multiagent in a distributed environment and leaming results can be shared by a group of design agents. This system has partly been implemented in Visual C++ based on Mechanical Desktop 2.0 environment.
