# Spatial order increases cooperation in quantum prisoner's dilemma game

##### February 25, 2014 by Lisa Zyga feature

Although we know that cooperating with others usually brings about the best results for everyone, oftentimes being uncooperative can be more beneficial to an individual who takes advantage of others who do cooperate. One of the most well-known tools for investigating cooperation is the prisoner's dilemma game, in which players receive different pay-offs depending on whether they and their partner decide to cooperate or not. In this game, a player receives the highest pay-off when she chooses to defect while her partner all-too-trustingly cooperates, and consequently receives the lowest pay-off.

In a new study, Ramón Alonso-Sanz, a researcher at the Technical University of Madrid, has developed and analyzed a spatial formulation of a quantum version of the prisoner's dilemma game. He found that spatial order increases cooperation compared with the original version of the game.

"Cooperation is not a 'rational' choice in the simple prisoner's dilemma game," Alonso-Sanz told Phys.org. "Thus, whatever mechanism that may help to explain the persistence of cooperation is always celebrated."

The game is played on a grid—or "cellular automaton"—with multiple players. There are two types of players, A and B, who alternately occupy the spaces of the grid like white and black spaces on a checkerboard.

Each player interacts only with his four adjacent partners (up, down, left, and right), all of whom are the opposite type of player as him. In each round, a player's pay-off is the sum of the pay-offs from these four interactions. (Pay-offs are 5, 4, 2, and 1 for DC, CC, DD, and CD, respectively, where C = cooperation and D = defection, with the first letter representing the player's choice and the second letter representing the choice of his partner.)

In the next round, the player adopts the strategy used by the most successful of the four players in the spaces diagonal to him (all of whom are the same type as him) or his own strategy—whichever received the highest pay-off in the previous round. Following this "imitation of the best" strategy, the game evolves over multiple iterations.

When Alonso-Sanz performed classical simulations of this spatial prisoner's dilemma game, he found that defection spread quickly. Even when beginning the game with a scenario where all but one A player chose to cooperate, the results showed that that one player's choice to defect spread to all of the other A players. In a 6x6 grid, every A player chose defection after only four iterations of the game.

But the game evolves very differently when is incorporated. When players can use quantum strategies, a player's defective choice does not spread through the game space, even though it may offer the highest payoff. This is because quantum entanglement changes the probabilities of the pay-offs. Whereas defection spreads in the classical version of the game, in the quantum version defection actually fades and instead mutual cooperation among all players emerges.

"The parameter pay-offs (e.g., 5,4,2,1) are of course unaltered," Alonso-Sanz explained. "What varies are the accessible probabilities, and consequently the mean (or expected) payoffs."

Although intriguing, this difference between the classical and quantum versions of the prisoner's dilemma game is not unexpected. In 1999, researchers found that classical and quantum formulations of the prisoner's dilemma game have different pairs of strategies that are in Nash equilibrium, meaning that the strategies of both players are the best response to one another. In the classical version, mutual defection is the only pair of strategies in Nash equilibrium. Even though both players would receive a larger payoff if they both cooperated, mutual cooperation is not a pair of strategies in Nash equilibrium because either player could do better for herself by choosing to defect.

However, in the quantum version, mutual cooperation is in Nash equilibrium. That is, when two players both choose cooperation, neither player can do better by choosing to defect instead. The current study demonstrates how this equilibrium easily emerges in a spatial formulation of the prisoner's dilemma game.

By making some modifications to the game, Alonso-Sanz found some other interesting results. In an unfair situation in which A-type players can use quantum strategies but B-type players are restricted to classical strategies, the A players rapidly take advantage of the B players. However, when the quantum A players are given a memory and forced to take into account the outcomes of several previous iterations, they actually choose worse strategies than the classical B players. This game results in both kinds of players receiving lower pay-offs than they would otherwise receive.

Overall, the results here are similar to the results from Alonso-Sanz's previous work on a game called Battle of the Sexes, which deals with coordination. As in the case of the prisoner's dilemma, a quantum formulation of the spatial Battle of the Sexes game evolves in a manner notably distinct from that of its classical counterpart: specifically, it leads to a more homogeneous distribution of coordination rather than clusters of coordination. Within the context of these spatially formulated quantum games, researchers may better understand how cooperation and coordination materialize in a society of competitive individuals.

More information: Ramón Alonso-Sanz. "A quantum prisoner's dilemma cellular automaton." Proceedings of The Royal Society A. DOI: 10.1098/rspa.2013.0793

Ramón Alonso-Sanz. "On a three-parameter quantum battle of the sexes cellular automaton." Quantum Information Processing. DOI: 10.1007/s11128-012-0496-2

Ramón Alonso-Sanz. "A quantum battle of the sexes cellular automaton." Proceedings of The Royal Society A. DOI: 10.1098/rspa.2012.0161

## Related Stories

#### To build a cooperative society, is it better to punish or reward?

April 19, 2010

(PhysOrg.com) -- One of the basic components of a functional, cooperative society is a code of law, where the laws are usually enforced by some kind of incentive. Social incentives can either be positive (rewards) or negative ...

#### Quantum strategy offers game-winning advantages, even without entanglement

March 13, 2012

(PhysOrg.com) -- Quantum correlations have well-known advantages in areas such as communication, computing, and cryptography, and recently physicists have discovered that they may help players competing in zero-sum games, ...

#### Link between quantum physics and game theory found

July 12, 2013

(Phys.org) —A deep link between two seemingly unconnected areas of modern science has been discovered by researchers from the Universities of Bristol and Geneva.

#### Generosity leads to evolutionary success, biologists show

September 2, 2013

With new insights into the classical game theory match-up known as the "Prisoner's Dilemma," University of Pennsylvania biologists offer a mathematically based explanation for why cooperation and generosity have evolved in ...

#### Study suggest people act fairly due to spite, not altruism

February 12, 2014

(Phys.org) —A study done by philosophers Patrick Forber of Tufts University and Rory Smead of Northwestern University, suggests fairness in societies evolves out of a fear of spite from others, rather than due to an increase ...

#### Hit video game 'The Last of Us' gets new chapter

February 15, 2014

Sony Computer Entertainment America added a new chapter Friday to popular video game "The Last of Us," filled with combat against zombie-like infected people and tailored for PlayStation 3 consoles.

## Recommended for you

#### Engineers quantify amount of Android root exploits available in commercial software

October 13, 2015

In recent years the practice of Android rooting, that is the process of allowing an Android phone or tablet to bypass restrictions set by carriers, operating systems or hardware manufacturers, has become increasingly popular.

#### Should cars be fully driverless? No, says an engineer and historian

October 13, 2015

If you follow technology news—or even if you don't—you have probably heard that numerous companies have been trying to develop driverless cars for a decade or more. These fully automated vehicles could potentially be ...

#### With this new universal wireless charger, compatibility won't be an issue

October 13, 2015

A wireless charger that's compatible with different consumer electronics from different brands is one step closer to becoming a reality thanks to research by electrical engineers at the University of California, San Diego.

#### New research method identifies stealth attacks on complicated computer systems

October 13, 2015

Three Virginia Tech computer scientists are unveiling a novel approach to discovering stealth attacks on computers at the annual ACM Conference on Computer and Communications Security.

#### Toyota promises better mileage and ride with Prius hybrid

October 13, 2015

Toyota Motor Corp. released details for its fourth-generation Prius on Tuesday, promising that improvements in the battery, engine, wind resistance and weight mean better mileage for the world's top-selling hybrid car.

#### Vehicle fuel economy standards as global climate policy: How much can they deliver and at what cost?

October 12, 2015

Over the past decade, many countries and regions seeking to reduce climate-warming carbon dioxide emissions have adopted more aggressive fuel economy standards designed to boost the efficiency of new, light-duty cars and ...

##### Infinum
5 / 5 (1) Feb 25, 2014
A whole article about entanglement and quantum prisoners' dilemma but not one word about the exact difference of the algorithms used.

So what are those quantum modifications to the classical rules of the game exactly?
##### antialias_physorg
not rated yet Feb 25, 2014
See this part
This is because quantum entanglement changes the probabilities of the pay-offs. Whereas defection spreads in the classical version of the game, in the quantum version defection actually fades and instead mutual cooperation among all players emerges.
"The parameter pay-offs (e.g., 5,4,2,1) are of course unaltered," Alonso-Sanz explained. "What varies are the accessible probabilities, and consequently the mean (or expected) payoffs."

Also see the second to last paragraph.