Metaheuristic entry points for harnessing human computation in mainstream games

Abstract

In this work, we describe a promising approach to harnessing human computation in mainstream video games. Our hypothesis is that one of the best approaches to seamlessly incorporating harnessing withing these games is by examining existing game mechanics and matching them to meta-heuristic algorithms. In particular, we believe that the best choices for early exploration of this problem are nature inspired meta-heuristic algorithms for combinatorial optimization problems. In this paper, we will describe the problem in more detail and describe two proof of concept games that demonstrate the viability of this approach. The first game is designed to be incorporated in Real-time Strategy games within the resource gathering aspects of these games, and the algorithm and problem that are used is related to Ant Colony Optimization and the Traveling Salesman Problem. The second game explores a racing game where the problem and algorithm are embedded in the numerical characteristics of the racer such as speed, agility, and jump power. These characteristics represent current solutions to different traveling salesman problems, and the solutions are modified through training and mating of racers; this is analogous to mutations and crossbreeding in genetic algorithms. © 2013 Springer-Verlag Berlin Heidelberg.