Iterated Prisoner's Dilemma Simulation (IPDS)

Team members:

• Mina Krivokuća
  • Index: 215/2016
  • E-mail: [email protected]
• Darinka Zobenica
  • Index: 75/2016
  • E-mail: [email protected]
• Vladimir Batoćanin
  • Index: 74/2016
  • E-mail: [email protected]

Libraries used:

• qCustomPlot
• Qt
• Catch2

ENG

IPDS is an interactive simulation of a widely studied gametheoretical concept - Prisoner's Dilemma.

Prisoner's Dilemma is a situation wherein two people have a choice between cooperation or betrayal, each yielding them a certain award or punishment. In the traditional formulation, if two criminals are caught by the police who is going to convict them on a lesser charge, but have no proof for the more serious one:

• if they both choose to stay silent and don't betray one another, they both get a short prison sentence
• if one makes a bargain with the police and betrays the other they get away without a sentence, while the other person gets a really long sentence
• if they both betray each other they both get medium-length sentences.

In this formulation, knowing nothing about what the other person might do, the smartest thing to do is to always betray them. But what becomes of this concept when we include the possibility of multiple rounds of the same game, and what if the participants can remember other people and decide what to do based on their previous interactions? This creates a complex almost life-like system that can inform our views of evolution, biology, economics, and other real-world phenomena.

Our simulation aims to provide an interactive interface for the users to test out their models. They'll be able to choose the kinds of strategies the specimen in the simulation will use, how many specimen will be using each strategy, etc.

---

SRB

IPDS je interaktivna simulacija detaljno izučavanog koncepta iz teorije igara - zatvorenikove dileme.

Zatvorenikova dilema je situacija u kojoj dve osobe imaju izbor da sarađuju ili ne, i gde na osnovu tog izbora dobijaju određenu nagradu ili kaznu. U tradicionalnoj formulaciji, ako su dva kriminalca uhapšena i policija ima osnova da ih osudi za manji zločin, ali im nedostaju dokazi za ozbiljniji zločin:

• ako oba kriminalca prećute i ne izdaju jedan drugog, dobijaju kratku zatvorsku kaznu
• ako se jedan dogovori s policijom ne dobija nikakvu kaznu u zamenu za to da svedoči protiv kolege, koji onda dobija jako dugu kaznu
• ako obojica izdaju jedan drugog, obojica dobijaju srednje dugu kaznu

U ovoj formulaciji, ukoliko zaista ne znaju ništa o tome šta će onaj drugi uraditi, najpametniji izbor je uvek izdati drugu osobu. Ali šta se sa ovim konceptom dešava ako dozvolimo da se igra više rundi iste igre? Ovo stvara kompleksan sistem koji imitira razne sisteme koje vidimo u stvarnom svetu: evoluciju, biologiju, ekonomiju...

Cilj naše simulacije je da napravimo interaktivan interfejs preko kog korisnici mogu da testiraju svoje modele ove dileme. Moći će da izaberu strategije koje će jedinke u simulaciji koristiti, koliko će jedinki koristiti koju simulaciju, itd.