| Summary: | 碩士 === 國立臺南大學 === 資訊工程學系碩士班 === 99 === Information technology is growing fast in the twenty-first century. One of the important goals that the computer scientists try their best to reach is to apply the artificial intelligence to the intelligent systems. Human pursues the growth of the knowledge and attempts to reach the creator’s stature. Nowadays, creating the machines with the intelligence is still quite difficult; however, this will not affect the computer scientists’ great enthusiasm to explorer the machines with intelligence. This thesis tries to (1) combine the technologies of the fuzzy ontology, evolutionary learning, and fuzzy markup language to make the machines have the similar learning and memorizing abilities to humans, and to (2) establish the architecture of the intelligent system to apply it to the Go and NoGo games. Based on the knowledge base and rule base constructed by humans, the experimental results show that there is a good development when we combine the technologies of the fuzzy ontology with the fuzzy inference mechanism to apply to the computer games. Additionally, the experimental results also indicate that the intelligent computer game system can learn the membership functions according to the past experience and implemented results via the proposed genetic learning mechanism to achieve the goal of continuously evolutionary reinforcement learning. It is believed that integrating the fuzzy ontology with the evolutionary learning will be a highly potential research direction for different kinds of the computer games. In the future, the intelligent system with the self-learning ability will gradually become mature to allow the machines to be much closer to the human’s life and to have much more similar thinking, reasoning, and learning abilities to the human.
|
|---|
