Chain of command in autonomous cooperative agents for battles in real-time strategy games

Damon Daylamani-Zad,Ioannis Th Paraskevopoulos,Letitia B Graham

doi:10.1007/s40692-018-0119-8

Damon Daylamani-Zad, Ioannis Th Paraskevopoulos + Show 1 more

Open Access

https://doi.org/10.1007/s40692-018-0119-8

Copy DOI

Abstract

This paper investigates incorporating chain of command in swarm intelligence of honey bees to create groups of ranked co-operative autonomous agents for an RTS game in to create and re-enact battle simulations. The behaviour of the agents are based on the foraging and defensive behaviours of honey bees, adapted to a human environment. The chain of command is implemented using a hierarchical decision model. The groups consist of multiple model-based reflex agents, with individual blackboards for working memory, with a colony level blackboard to mimic the foraging patterns and include commands received from ranking agents. An agent architecture and environment are proposed that allows for creation of autonomous cooperative agents. The behaviour of agents is then evaluated both mathematically and empirically using an adaptation of anytime universal intelligence test and agent believability metric.

Highlights

Most genres of video games require a degree of artificial intelligence (AI) either as support with progress or as opponents
This paper investigates incorporating chain of command in swarm intelligence of honey bees to create groups of ranked co-operative autonomous agents for an real-time strategy (RTS) game in to create and re-enact battle simulations
This research investigated the idea of incorporating chain of command with swarm intelligence of honey bees when foraging and defending their nests to create a group of co-operative agents with leadership and tactical decision-making

Summary

Introduction

Most genres of video games require a degree of artificial intelligence (AI) either as support with progress or as opponents. These could be central intelligences, individual agents or groups of agents. Specific genres of games require large numbers of individual AI units to work together for or against the player, such as in real-time strategy (RTS) or Tower Defence games. Multi-agent approaches and swarm intelligence are inspired on the ability of social animals and crowds to work together as a group without the need for a leader to delegate tasks. Individuals in a swarm are unable to find a solution to a colony’s problems alone; by interacting with each other and making decisions based on local information, they can find a solution at the colony level (Garnier et al 2007)

Objectives

Methods

Results

Conclusion