Mode of access: World Wide Web.

System requirements: Adobe Acrobat Reader.

Part of: Synthesis digital library of engineering and computer science.

Includes bibliographical references (pages 161-170).

1. Introduction -- 1.1 Prisoner's dilemma and other games -- 1.2 Digital evolution: evolving game players -- 1.2.1 Optimization vs. co-evolution -- 1.3 From checkers through the rise of the go machines -- 1.4 Why are the simple games so hard? --

2. The first place where trouble arose -- 2.1 Representation: what is it? -- 2.1.1 Finite state machines, direct representation -- 2.1.2 Finite state machines with a cellular representation -- 2.1.3 Boolean formulas -- 2.1.4 Function stacks -- 2.1.5 ISAc lists -- 2.1.6 Markov chains and look-up tables -- 2.1.7 Artificial neural nets -- 2.2 Fingerprinting: a tool for comparing across representations -- 2.2.1 Definition of fingerprints -- 2.2.2 Example fingerprint computation -- 2.2.3 Fingerprint results -- 2.3 The uncontrolled variable discovery -- 2.3.1 Experimental design -- 2.3.2 Experimental results -- 2.3.3 Conclusion --

3. Problems beyond representation -- 3.1 Changing the payoff matrix changes the agents that arise -- 3.1.1 Design of experiments -- 3.1.2 Results and discussion -- 3.1.3 Conclusions -- 3.2 Changing the level of resources -- 3.2.1 Design of experiments -- 3.2.2 Results and discussion -- 3.2.3 Conclusions and next steps -- 3.3 Algorithm details matter to population size, elite fraction, and mutation rate -- 3.3.1 Design of experiments -- 3.3.2 Results and discussion -- 3.3.3 Conclusions and next steps -- 3.4 Including tags and geography -- 3.4.1 Grid-based nIPD and agent specifications -- 3.4.2 Experimental design -- 3.4.3 Results -- 3.4.4 Conclusions --

4. Does all this happen outside of prisoner's dilemma? -- 4.1 Coordination prisoner's dilemma, rock-paper-scissors, and morphs -- 4.1.1 Design of experiments -- 4.1.2 Results and discussion -- 4.1.3 Conclusions -- 4.2 Divide-the-dollar--a more complex game -- 4.2.1 Generalize divide-the-dollar -- 4.2.2 Design of experiments -- 4.2.3 Results and discussion -- 4.2.4 Conclusions and next steps -- 4.3 The snowdrift game -- 4.3.1 The game models -- 4.3.2 Experimental results -- 4.3.3 Discussion -- 4.3.4 Conclusions --

5. Noise! -- 5.1 Noisy games are different -- 5.1.1 Experimental design -- 5.1.2 Analysis techniques -- 5.1.3 Results and discussion -- 5.1.4 Conclusions and next steps -- 5.2 Evolutionary velocity -- 5.2.1 Definition of evolutionary velocity -- 5.2.2 Analysis of evolutionary velocity --

6. Describing and designing representations -- 6.1 Does the representation have internal states? -- 6.2 Does the representation use external state information? -- 6.3 Can the representation learn? -- 6.4 Does the representation use random numbers? -- 6.5 Complex representations --

Bibliography -- Authors' biographies.

Abstract freely available; full-text restricted to subscribers or individual document purchasers.

Compendex

INSPEC

Google scholar

Google book search

Evolving agents to play games is a promising technology. It can provide entertaining opponents for games like Chess or Checkers, matched to a human opponent as an alternative to the perfect and unbeatable opponents embodied by current artificial intelligences. Evolved agents also permit us to explore the strategy space of mathematical games like Prisoner's Dilemma and Rock-Paper-Scissors. This book summarizes, explores, and extends recent work showing that there are many unsuspected factors that must be controlled in order to create a plausible or useful set of agents for modeling cooperation and conflict, deal making, or other social behaviors. The book also provides a proposal for an agent training protocol that is intended as a step toward being able to train humaniform agents - in other words, agents that plausibly model human behavior.

Also available in print.

Title from PDF title page (viewed on August 23, 2017).