What is an Agent? Lecture Notes

What is an Agent? Agents & Environments. Agent Characterisation We have 6 different definitions for the Agent. 1. Agent Definition - a agent is an entity which is:  situated in some environment  autonomous: means it can work by itself, no need for human intervention or other...

What is an Agent? Agents & Environments. Agent Characterisation We have 6 different definitions for the Agent. 1. Agent Definition - a agent is an entity which is:  situated in some environment  autonomous: means it can work by itself, no need for human intervention or other software processes, it controls its own actions  flexible: responsive for his actions, proactive and social 2. Agent Definition - agent: one that acts or has the power or authority to act or represent another. 3. Agent Definition - an agent is anything that can be viewed as perceiving its environment through sensors and acting upon the environment through effectors. 4. Agent Definition - autonomous agents are computer programs or systems that operate on their own in a changing environment - they can observe whats happening around them, make decisions and take actions without needing direct human control 5. Agent Definition - intelligent agents are software programs that perform tasks for users or other programs - they operate with some level of autonomy and use knowledge and they try to understand the user goals to make decisions 6. Agent Definition -an agent is a computer program that works on its own and can communicate with other systems or programs to perform specific tasks in an application. The agent takes sensory input from its environment and produces as output actions that affect it. We have internal and external Environments of an Agent: 1. internal: goals, sensors, beliefs, profile, knowledge, effectors, abilities etc. 2. external: user, other humans, other agents, platforms, servers, networks etc. Autonomous agents Biological Agents Robotic Agents Computational Agents Software Agents Artificial Life Agents Task-Specific Agents Entertainment Agents Viruses *In this class we are gonna be more focused in these in the red highlight. An agent is responsible for satisfying specific goals. There can be different types of goals such as achieving a specific status, maximising a given function etc. Thestate of an agent includes state of its internal environment + state of knowledge and beliefs about its external environment. Thestate of an agent includes state of its internal environment + state of knowledge and beliefs about its external environment. An agent is situated in an environment, which can be:  Accessible / partially accessible / inaccessible.  Deterministic / nondeterministic  Static / dynamic In complex environments: An agent do not have complete control over its environment, it just have partial control. Partial control means that an agent can influence the environment with its actions. Key problem for an agent: deciding which of its actions it should perform in order to best satisfy its design objectives. Sensor input Agent Action output Environm ent History History represents the interaction between an agent and its environment. Standard Agents A Standard agent decides what action to perform on the basis of his history (experiences). Purely Reactive Agents A purely reactive agent decides what to do without reference to its history (no references to the past).  The agent starts in some internal state, then observes its environment state. The internal state of the agent is updated with next(i0,see(s)). The action selected by the agent becomes action(next(i0,see(s))), and it is performed. The agent repeats the cycle observing the environment. Agents actions can be direct and indirect. Direct: the affect properties of the objects in the environment. Indirect: send messages with the aim of affecting mental attitudes of other agents. Multi-Agent Systems (MAS) The MAS means that a cooperative work environment is made up of different software components that work together. By collaborating these components can handle complex problems more effectively than if they worked alone. Three main approaches:  Cooperative interaction  Contract-based co-operation  Negotiated cooperation The Mobile Agent is the entity that moves between platforms.