University of Twente Student Theses

Login

Computations in Stochastic Game Theory : large sets of rewards in games with communicating states and frequency-dependent transition probabilities and stage payoffs

Samuel, L.N. (2017) Computations in Stochastic Game Theory : large sets of rewards in games with communicating states and frequency-dependent transition probabilities and stage payoffs.

[img]
Preview
PDF
1MB
Abstract:With the amount of conceptual literature on 2-person social dilemmas, this thesis takes the necessary step of creating an algorithm that is able to compute what has been modeled. That is: dynamic 2-state, 2-player, 2-action stochastic competitive games with allowances for frequency-dependent transition probabilities and frequency-dependent stage payoffs. To demonstrate its usage, I employ a ‘Commons’-type social dilemma where 2 players compete over a renewable common-pool resource. In this example, both the transition probabilities and stage payoffs possess frequencydependencies that are linear. Note that adherence to such a limitation is not necessary for the stage payoffs, but may be for the transition probabilities. Results in both the repeated game model and the two stochastic models are in line with results found in prior studies. That is, frequency-dependent stage payoffs cause a reduction in the value of the payoffs available as the reward set heads south-west in R2. More noteworthy, results in the stochastic models with frequency-dependent stage payoffs showthat frequency-dependent transition probabilities have significant reduction effects on the probability of obtaining a higher payoff for either player. The reward sets in both situations increase in size by stretching toward the origin, compared to the frequency-independent stage payoffs reward sets. This increase is compensated in part through a narrowing of the set’s shape in R2 space which results in smaller differences between both players when one gains more than the other.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:54 computer science
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:http://purl.utwente.nl/essays/72328
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page