Group Members

IT23689176-Tharuka HMND

IT23657700-Sandil Helitha Perera

It23651388-S.I.Rathnayaka

Agent Types & Architectures

Introduction

AI agents are the autonomous systems perceive its environment and make decisions based on the environmental behaviors. The architectures of the most agents simply relies on 4 main modules. Profiling module(to grasp the perception or role of the agent), memory module(to store the past experiences of the agent), planning module(use to plan the tasks efficiently for the users), Action module(use to initiate the plan efficiently).

Core explanation

Core explanation mainly emphasized on various types of design patterns, reasoning strategies and execution of the autonomous actions. Mostly agents have the capability of vertical layering this has the capability to separate routines to different layers. This will increase the efficiency and speedup. There are several components that will contribute to this process;

1. Behavior based component - Handles immediate reactions and routine tasks
2. Agent Type Equivalent – This will operates on external facts and information’s
3. Plane based component – Goal directed planning.
4. Corporation component – manages interactions with other agents which is integrated to the current agent.

Control flow

This is the feature that agent switches between simple to complex tasks.

• Bottom up approach(competent driven) – If Behaviour based component identifies an obstacle it will immediately reacted.

Behaviour based component -> Plane based component -> Corporation component

• Top down approach – If a highr layer identifies a solution it will transfer to the lower layers