A game, known as “Colonel Blotto,” has been used to analyse the potential outcomes of elections and other similar two-party conflicts since 1921 but it has been limited by its lack of a definitive solution.
Now, computer scientists from the University of Maryland, Stanford University and Microsoft Research have for the first time solved a game theory scenario that has given researchers headache for decades.
They developed an algorithm capable of solving the Colonel Blotto scenario. The algorithm could also provide political strategists, business leaders and other decision-makers with a powerful new tool for making informed choices.
Colonel Blotto pits two competitors against one another and requires each to make difficult decisions on how to deploy limited resources. Each player assigns a limited number of resources, or troops, to a number of battlefields, without the knowledge of their opponent’s strategy. Players win a given battlefield if they allocate more troops than their opponent; the player who wins the most battlefields also wins the game.
In real world, the game finds its use in scenarios such as a U.S. presidential general election. In this example, each candidate is a player; resources such as campaign staff, stump time and funding are the troops; and each state is a battlefield. The game can also apply to high-profile consumer product competition, such as the ongoing battle between Apple’s iPhone and Google’s Android mobile phone products.
Although the algorithm was limited by the large variety of possible the team overcame this by limiting the total number of possible strategies to a relative handful of representative choices.
This solution enabled the team to develop a generalized algorithm, which can now be applied to specific scenarios, such as the 2016 presidential election.
Reference: https://www.sciencedaily.com/releases/2016/02/160211190010.htm
seangametheory 