Complementing Solutions to Optimization Problems via Crowdsourcing on Video Game Plays

Abstract

Leveraging human insight and intuition has been identified as having the potential for the improvement of traditional algorithmic methods. For example, in a video game, a user may not only be entertained but may also be challenged to beat the score of another player; additionally, the user can learn complicated concepts, such as multi-objective optimization, with two or more conflicting objectives. Traditional methods, including Tabu search and genetic algorithms, require substantial computational time and resources to find solutions to multi-objective optimization problems (MOPs). In this paper, we report on the use of video games as a way to gather novel solutions to optimization problems. We hypothesize that humans may find solutions that complement those found mechanically either because the computer algorithm did not find a solution or because the solution provided by the crowdsourcing of video games approach is better. We model two different video games (one for the facility location problem and one for scheduling problems), we demonstrate that the solution space obtained by a computer algorithm can be extended or improved by crowdsourcing novel solutions found by humans playing a video game.