Background and Motivation of This Research on Artificial Intelligence

The emergence of computers along with the field of Artificial Intelligence (AI) in the recent past has had a significant impact on science and technology. These technologies have made a great impact on our life and our society. They could even be considered as having had the most significant influence on any technology during any period of history. Software engineering has evolved from traditional top-down, sequential programs through shared services, agents, and new tools used to insulate humans from mundane or dangerous tasks. 

Advances in computer technology have created super-fast computers with high-quality graphic cards and improved data storage, bandwidth, and data transaction speed. These improvements have stimulated the emergence of virtual environments with rich life-like graphics that simulate the physical world in great detail. 

The seamless ubiquity of the digital domain and fast Internet connection has created an environment in which information evolves. While people once elected whether or not to go online, today most remain connected to the environment and choose to engage with others in a digital community. These technological advancements provide the tools and platforms to develop, test and implement intelligent systems. 

Many researchers have focused on developing Artificial Intelligence theories and techniques to solve problems in a number of domains. Such techniques started from conventional approaches, which include searching and optimisation techniques, case-based and rule-based Artificial Intelligence, Finite State Machines (FSMs), and Fuzzy Logic [1]. 

Traditional AI techniques have evolved over the recent past to optimise solutions in different domains including searching, navigation, and pathfinding. For example, in the domains of navigation and pathfinding many graph-search algorithms were developed to perform efficiently in multiple situations. The A* algorithm [2] is one such search technique that is highly efficient in pathfinding. 

Many Artificial Intelligence techniques have evolved over the past decades to model cognitive aspects of human behaviour including perception, reasoning, communication, and learning. More advanced AI techniques were developed including Artificial Neural Networks (ANNs) and Evolutionary Algorithms (EAs), which were inspired by how the brain works and how biological life evolved respectively. One example of these techniques is the real-time NEAT [3], which combines the concepts of Artificial Neural Networks and Evolutionary Algorithms to improve agents’ behaviour. 

Automation has become necessary in various applications in today’s world where human intervention needs to be minimised. Agent-oriented technology is growing rapidly to enhance automation. Intelligent Agents are used to acting autonomously on behalf of humans in order to perceive the environment, make decisions, and plan a series of tasks to achieve certain goals [4]. Autonomy is a fundamental property of Intelligent Agents. Agents should also be able to communicate and interact with other Intelligent Agents in Multi-Agent System to perform sophisticated tasks that require the cooperation of disparate entities with dedicated capabilities. 

Using communication techniques, Intelligent Agents can exchange information and knowledge with all components within the environment. In Multi-Agent System, Intelligent Agents can break down tasks into subtasks and allocate each subtask to an Intelligent Agent that has the required capabilities. They can cooperate as a team to solve problems and achieve mutual goals. 

Artificial Intelligence techniques have significant potential applications in many fields; including disaster recovery, traffic control, space exploration, business management, and computer games. The ultimate goal of Artificial Intelligence is to create autonomous entities that are capable of acting reliably on behalf of humans. Such systems would have great advantages in domains where human intervention needs to be minimised. Dreams of creating autonomous systems have motivated me to research existing AI techniques and integrate and hybridise some of the AI techniques in Multi-Agent Systems to achieve better autonomy. Therefore, the aims of this research are to improve the performance and efficiency of Multi-Agent Systems in domains where agent teaming is greatly advantageous.

References and Credits

1. J. Haugeland. Artificial Intelligence: The Very Idea. MIT Press, Cam- bridge, MA, 1985.
2. P. E. Hart,  N. J. Nilsson,  and B. Raphael.  A formal basis for the heuristic determination of minimum cost paths. IEEE Transactions on Systems Science and Cybernetics SSC4, 4(2):100–107, July 1968.
3. K. O. Stanley, B. D. Bryant, and R. Miikkulainen. Real-time neuroevolution in the NERO video game. IEEE Transactions on Evolutionary Computation, 9:653–668, 2005.
4. M. Wooldridge and N. R. Jennings. Intelligent agents: Theory and practice. Knowledge Engineering Review, Cambridge [England]; New York, NY: Cambridge University Press, 10(2):115 – 152, 1995.