Motivation for Multi-Agents Collaboration

An algorithm running on a computer is limited by its memory and computation capacity. Though increasing the capacity of a computer may improve algorithms’s performance but generally increasing capacity is costly. If an algorithm’s performance can scale with adding computers, then it may be a cost effective way to achieve desired performance. These new algorithms will have to communicate their information to other computers running similar algorithm.

Even with all the available capacity, single agent algorithm may still take impractical amount of time to achieve desired result on large state space task. A well coordinated multi-agents may explore different parts of state space to yield better performance.

Tasks like forest survelliance (or wide range survelliance) require multiple agents to collaborate.