Cooperation is an important area of investigation for behavior analysis. The prisoner’s dilemma game (PDG) provides a useful scenario for studying cooperation in a behavior analytic paradigm. The PDG can be coupled with the concept of the metacontingency to investigate how various contingency arrangements support and promote cooperation in a group. Players in this experiment participated in a PDG and, in some conditions, were given the ability to fine other players but could not talk. The goal of this experiment was to investigate how players’ ability to fine one another affected the players’ patterns of cooperation, and whether fining itself was affected by the addition of a shared group consequence. The data show that participants cooperated in some conditions, but the fines did not seem to affect players’ rates of cooperation.

Recent research has shown that variation in the gene encoding for the oxytocin receptor (OXTR) contributes to individual differences in social-cognitive and emotional functioning in both clinical and non-clinical populations. OXTR has been associated with prosocial behavior, positive and negative emotionality, empathy, maternal sensitivity, and stress reactivity. To date, no study has investigated OXTR in the context of behavioral and emotional reactions to betrayals in trust. The present study examined how variation in 10 SNPs on OXTR may contribute to individual differences in behavior, emotional reactions, and perceptions following a betrayal in trust in an iterated prisoner’s dilemma game. Following correction for multiple testing, one SNP (rs237887) and two haplotypes (A-rs237887, C-rs2268490; G-rs237887, C-rs2268490) were significantly associated with positive emotional reactions to betrayal. In addition, one haplotype (C-rs9840864, T-rs2268490) was significantly associated with negative emotional reactions to betrayal. The present findings suggest that variation on OXTR may contribute to individual differences in emotional reactions to betrayals in trust.

O estudo de Teoria de Jogos tem se expandido para diversas áreas, tendo sua aplicação inicial na economia, hoje é utilizado na psicologia, na filosofia e tem um papel importantíssimo na biologia evolutiva. O seu sucesso está ligado ao fato de que os jogos têm o poder de prever interações usando conceitos simples como a cooperação e a competição. Dentre os jogos há o famoso de Dilema do Prisioneiro (PD), em que indivíduos completamente racionais devem optar entre cooperar ou trair (desertar) seu companheiro de jogo. A estratégia dominante e o equilíbrio de Nash, para o PD, é a deserção mútua visto que os indivíduos são sempre tentados a não cooperar. O dilema é que eles obteriam um ganho melhor se cooperassem mutuamente. Na vida real os indivíduos se encontram em várias situações nas quais eles devem optar entre ser egoístas ou altruístas e, frequentemente, acabam optando pelo altruísmo. Mesmo com a previsão da deserção na teoria clássica dos jogos, em 1992 Nowak e May (NOWAK; MAY, 1992) mostraram que cooperação é mantida em jogos com interação espacial e evolutivos A partir dessa descoberta, estudos de jogos em diversos tipos de rede foram propostos, entre eles as redes diluídas (que possuem sítios vacantes). Nesse tipo de rede foi observado que certas densidades favorecem a cooperação, particularmente próximo ao limiar de percolação para regras de atualização estocásticas (com ruído). Porém a probabilidade de troca do Replicador, mesmo sendo estocástica, não se encaixa nesse padrão observado. Descobrimos que esse comportamento anômalo está relacionado com estruturas formadas entre buracos e desertores que impedem alguns indivíduos de ter acesso ao ruído, assim a informação não flui livremente na rede. Consequentemente o sistema fica preso em um estado congelado, que pode ser quebrado com algum tipo de perturbação. Também abordamos a relação entre o limiar de percolação por sítio e a cooperação de uma forma mais quantitativa do que já foi apresentada até então, acompanhamos o desenvolvimento da cooperação dentro dos clusters e mostramos como o limiar de percolação afeta as estruturas básicas da rede. / The study of Game Theory, having its initial application in economics, has expanded to several areas and is now used in psychology, philosophy and plays a major role in evolutionary biology. Its success is related to the fact that games have the power to predict and study interactions using simple concepts such as cooperation and competition. Among the games there is the famous Prisoner Dilemma (PD), where completely rational individuals have to choose between cooperating or betraying their game partner. The dominant strategy and the Nash equilibrium for PD is mutual desertion as individuals are always tempted to not cooperate. The dilemma is that they would get a higher payoff if they mutually cooperated. In real life, individuals find themselves in various situations where they must choose to be selfish or altruistic, and often they choose altruism. Even with the prediction of defection in classical game theory, in 1992, Nowak and May (NOWAK; MAY, 1992) showed that cooperation is maintained in evolutionary spatial games. With this discovery, the study of games on several types of networks was proposed, among them the diluted networks (which have vacant sites) In this type of lattice, it was observed that at certain densities cooperation is promoted, particularly close to the percolation threshold for stochastic updating rules. However, the exchange probability of the Replicator dynamics, despite being stochastic, does not obey this observed pattern. We found that this anomalous behavior is related to structures formed between holes and defectors that prevent some individuals from having access to noise, so information does not flow freely in the network. Consequently the system becomes trapped in a frozen state, but this state can be broken by perturbing the system. We also address the relationship between the percolation threshold and cooperation in a more quantitative way than has been presented lately, by following the development of cooperation within clusters and showing how the percolation threshold affects the basic structures of the lattice.

This study got its origin in the failed climate negotiations in the Copenhagen 2009 summit. By conducting a public good game, with participants from China and Sweden, my study indicates that previous studies on public good games can predict the outcome of the game to a quit large extent even though most of my statistical tests came out statistically insignificant. My study also indicates that by framing the game as climate negotiations there were no statistical significant difference on the level of contributions in comparison to the unframed versions of the game. The awareness of the issues with emissions, global warming and other environmental problems are pretty high but even so when push comes to shove gains in the short run are prioritized to gains in the long run. There are however hypothetical willingness to come to term with the environmental issues. The results of the study indicate that the outcome of the Copenhagen summit can be avoidable but would need additional experiments made on cultural differences and behavior.

Separate essays on entrepreneurship, joint venturing, and trade comprise this thesis. The emergence of entrepreneurship is common in the real world but relatively less so in classical economic models. If industry incumbents are attributed with full rationality and perfect foresight, then there are few, if any, profitable opportunities left for new entrants (entrepreneurs) to exploit. This essay explains how entrepreneurs can emerge in a dynamic world when firms must choose between a technology strategy that is either statically or dynamically efficient. A model is developed which shows how such opportunities for new entry can occur when incumbents are caught in a Prisoners’ Dilemma game involving technology strategy. A relevance measure and policy implications are then explored. Joint ventures, especially of the R&D type, are becoming increasingly important as a way to gain needed technological and market competencies. Unfortunately, many joint ventures have the characteristics of a Prisoners’ Dilemma. Firms may cooperate or defect in the venture. If contracts, side-payments, and third-party verification of the venture outcome are unavailable, then the dominant solution to the Prisoners’ Dilemma (mutual defection) results. This paper proposes the use of an ex-ante auction to obtain a Pareto-improvement for these ventures. A Pareto-improvement is assured when non-transferable costs and benefits of firms are not conditional on joint venture strategies. When this condition is not met restrictions are required to obtain the Pareto-improvement. The problem of trade between countries that share an international open access resource is becoming significant as the world reaches the limits of critical shared resource stocks. It is modelled as a world with one primary factor, two intermediate goods, one final good (harvested from the open access resource), and two nations where it is assumed that either the trading takes place over one stage (nations are price-takers), or two stages (nations have market power). Imperfect competition and open access generated externalities affect the trading efficiency. To maximize world welfare this essay recommends subsidizing R&D where comparative advantage exists, and creating international agreements to ensure the one-stage game structure is used when trading.

O estudo de Teoria de Jogos tem se expandido para diversas áreas, tendo sua aplicação inicial na economia, hoje é utilizado na psicologia, na filosofia e tem um papel importantíssimo na biologia evolutiva. O seu sucesso está ligado ao fato de que os jogos têm o poder de prever interações usando conceitos simples como a cooperação e a competição. Dentre os jogos há o famoso de Dilema do Prisioneiro (PD), em que indivíduos completamente racionais devem optar entre cooperar ou trair (desertar) seu companheiro de jogo. A estratégia dominante e o equilíbrio de Nash, para o PD, é a deserção mútua visto que os indivíduos são sempre tentados a não cooperar. O dilema é que eles obteriam um ganho melhor se cooperassem mutuamente. Na vida real os indivíduos se encontram em várias situações nas quais eles devem optar entre ser egoístas ou altruístas e, frequentemente, acabam optando pelo altruísmo. Mesmo com a previsão da deserção na teoria clássica dos jogos, em 1992 Nowak e May (NOWAK; MAY, 1992) mostraram que cooperação é mantida em jogos com interação espacial e evolutivos A partir dessa descoberta, estudos de jogos em diversos tipos de rede foram propostos, entre eles as redes diluídas (que possuem sítios vacantes). Nesse tipo de rede foi observado que certas densidades favorecem a cooperação, particularmente próximo ao limiar de percolação para regras de atualização estocásticas (com ruído). Porém a probabilidade de troca do Replicador, mesmo sendo estocástica, não se encaixa nesse padrão observado. Descobrimos que esse comportamento anômalo está relacionado com estruturas formadas entre buracos e desertores que impedem alguns indivíduos de ter acesso ao ruído, assim a informação não flui livremente na rede. Consequentemente o sistema fica preso em um estado congelado, que pode ser quebrado com algum tipo de perturbação. Também abordamos a relação entre o limiar de percolação por sítio e a cooperação de uma forma mais quantitativa do que já foi apresentada até então, acompanhamos o desenvolvimento da cooperação dentro dos clusters e mostramos como o limiar de percolação afeta as estruturas básicas da rede. / The study of Game Theory, having its initial application in economics, has expanded to several areas and is now used in psychology, philosophy and plays a major role in evolutionary biology. Its success is related to the fact that games have the power to predict and study interactions using simple concepts such as cooperation and competition. Among the games there is the famous Prisoner Dilemma (PD), where completely rational individuals have to choose between cooperating or betraying their game partner. The dominant strategy and the Nash equilibrium for PD is mutual desertion as individuals are always tempted to not cooperate. The dilemma is that they would get a higher payoff if they mutually cooperated. In real life, individuals find themselves in various situations where they must choose to be selfish or altruistic, and often they choose altruism. Even with the prediction of defection in classical game theory, in 1992, Nowak and May (NOWAK; MAY, 1992) showed that cooperation is maintained in evolutionary spatial games. With this discovery, the study of games on several types of networks was proposed, among them the diluted networks (which have vacant sites) In this type of lattice, it was observed that at certain densities cooperation is promoted, particularly close to the percolation threshold for stochastic updating rules. However, the exchange probability of the Replicator dynamics, despite being stochastic, does not obey this observed pattern. We found that this anomalous behavior is related to structures formed between holes and defectors that prevent some individuals from having access to noise, so information does not flow freely in the network. Consequently the system becomes trapped in a frozen state, but this state can be broken by perturbing the system. We also address the relationship between the percolation threshold and cooperation in a more quantitative way than has been presented lately, by following the development of cooperation within clusters and showing how the percolation threshold affects the basic structures of the lattice.

Separate essays on entrepreneurship, joint venturing, and trade comprise this thesis. The emergence of entrepreneurship is common in the real world but relatively less so in classical economic models. If industry incumbents are attributed with full rationality and perfect foresight, then there are few, if any, profitable opportunities left for new entrants (entrepreneurs) to exploit. This essay explains how entrepreneurs can emerge in a dynamic world when firms must choose between a technology strategy that is either statically or dynamically efficient. A model is developed which shows how such opportunities for new entry can occur when incumbents are caught in a Prisoners’ Dilemma game involving technology strategy. A relevance measure and policy implications are then explored. Joint ventures, especially of the R&D type, are becoming increasingly important as a way to gain needed technological and market competencies. Unfortunately, many joint ventures have the characteristics of a Prisoners’ Dilemma. Firms may cooperate or defect in the venture. If contracts, side-payments, and third-party verification of the venture outcome are unavailable, then the dominant solution to the Prisoners’ Dilemma (mutual defection) results. This paper proposes the use of an ex-ante auction to obtain a Pareto-improvement for these ventures. A Pareto-improvement is assured when non-transferable costs and benefits of firms are not conditional on joint venture strategies. When this condition is not met restrictions are required to obtain the Pareto-improvement. The problem of trade between countries that share an international open access resource is becoming significant as the world reaches the limits of critical shared resource stocks. It is modelled as a world with one primary factor, two intermediate goods, one final good (harvested from the open access resource), and two nations where it is assumed that either the trading takes place over one stage (nations are price-takers), or two stages (nations have market power). Imperfect competition and open access generated externalities affect the trading efficiency. To maximize world welfare this essay recommends subsidizing R&D where comparative advantage exists, and creating international agreements to ensure the one-stage game structure is used when trading. / Business, Sauder School of / Graduate

According to game theory, dominant strategy of Prisoner’s Dilemma game is defecting. Since online trading between two strangers falls in the realm of a Prisoner’s Dilemma, failure should be the only logical conclusion of such electronic commerce situation. Traders are less likely to cooperate since they are not required to deal with the same buyer again given the enormous population of online traders. Thus one could argue that markets like eBay should never exist. Then what is the reason behind resounding success of such electronic markets? The answer lies in the reputation system that they established. Google page rankings and reader reviews on various sites like Amazon.com and Epinions.com are some other examples of the reputation system pointing towards the enormous influence that they wield in our everyday life. This research thus attempts to analyze importance of reputation system in an online trading using Agent Based Simulation. The effect of reputation system on the sustainability of mutual cooperation between online traders is analyzed by taking into account key factors like level of gullibility of online traders and the weight of influence given to their past behavior. Results indicate that reputation system is certainly capable of sustaining cooperation among online traders. However, weightage given to the past behavior of agents should be decided only after taking into account the level of gullibility of agents. Use of agent based modeling for studying dynamics of online trading is a relatively new concept. Sustaining cooperation among online traders is often perceived to be difficult task. The reputation system proposed in this research is expected to make contribution towards this goal.

Kooperatives Verhalten lässt sich in vielen Bereichen menschlichen Zusammenlebens sowie im gesamten Tierreich beobachten. In evolutionären Modellen wurde gezeigt, dass Netzwerkstrukturen die Kooperation erhöhen können. Empirische Studien versuchten vergeblich diesen Mechanismus auch bei Menschen nachzuweisen. Es scheint, als würden Netzwerke nur dann die Kooperation erhöhen, wenn die Strukturen nicht statisch sind, sondern dynamisch. Das heißt, dass die Individuen die Möglichkeit haben, ihre Partner zu wechseln. Eine wichtige – aber bislang unerforschte – Eigenschaft dynamischer Netzwerke ist jedoch, dass derartige Wechsel von Partnern in der Regel Kosten verursachen, ob in Form von Zeit oder Ressourcen. Kapitel I meiner Arbeit schließt diese Lücke, in dem es sich mit den Effekten von Kosten auf dynamischen Netzwerken befasst. Ich konnte nachweisen, dass Menschen seltener Interaktionen mit Partnern beendeten, wenn die Kontaktaufnahme mit einem neuen Partner mit Kosten verbunden war. Bei sehr hohen Kosten, wurden Partner so selten gewechselt, dass das Netzwerk fast statisch war. Interessanterweise blieb die Kooperation dennoch sehr hoch. Das bedeutet, dass für kooperatives Verhalten entscheidend ist, ob man die Möglichkeit hat, Partner zu wechseln. Im Gegensatz zu bisherigen Annahmen ist es daher nicht wichtig, wie oft tatsächlich Partner gewechselt werden, sondern lediglich ob es die Möglichkeit dazu gibt. In Kapitel II beschäftige ich mich mit optimalem Entscheidungsverhalten. Im sogenannten Judge-Advisor-System geht es darum, dass eine Person, der Judge, eine unbekannte numerische Größe schätzen will. Dazu erhält der Judge eine zweite unabhängige Schätzung als Rat von einer zweiten Person, des Advisor. Schließlich ist die Frage, wie der Judge optimal den Rat verwerten kann um seine Anfangsschätzung zu verbessern. Bisherige Forschung konzentrierte sich hauptsächlich auf zwei mögliche Methoden, (i) das Bilden des Mittelwerts und (ii) das Wählen der besseren Anfangsschätzung. Das Hauptargument für diese einfachen Methoden ist deren häufige Verwendung in bisherigen Experimenten. Allerdings wurden sehr wohl auch andere Gewichtungen beobachtet und daher ist eine gründliche Analyse der optimalen Gewichtung erforderlich. In der vorliegenden Arbeit leitete ich ein normatives Modell her, das beschreibt, unter welchen Bedingungen welche Methode das bestmögliche Ergebnis liefert. Es wurden drei Methoden verglichen: (i) das Bilden des Durchschnitts, (ii) das Wählen der besseren Anfangsschätzung, und (iii) das Bilden eines gewichtetet Mittelwerts, wobei das Gewicht vom Kompetenzunterschied abhängt. Welche Methode optimal ist, hängt davon ab, wie groß der Kompetenzunterschied ist und wie gut er vom Judge erkannt wird. Die Durchschnittbildung ist immer dann vorteilhaft, wenn der Kompetenzunterschied nicht groß ist, oder nur schwer richtig eingeschätzt werden kann. Wenig überraschend lohnt sich das Wählen der besseren Anfangsschätzung, wenn der Kompetenzunterschied hinreichend groß ist, vorausgesetzt es wird tatsächlich die bessere Anfangsschätzung gewählt. Wenn der Kompetenzunterschied vom Judge gut eingeschätzt werden kann, ist eine Entsprechende Gewichtung immer die beste Methode, unabhängig vom tatsächlichen Unterschied. In Übereinstimmung mit bisheriger Forschung wurde auch die Kombination von Durchschnittbildung und Wählen der besseren Anfangsschätzung untersucht. Diese Kombinationsmethode beruht darauf, bei als gering eingeschätztem Kompetenzunterschied den Durchschnitt zu bilden und ansonsten die bessere Anfangsschätzung zu wählen. Interessanterweise schneidet diese Kombinationsmethode sehr schlecht ab, was hauptsächlich daran liegt, dass zu oft die falsche Anfangsschätzung genommen würde. Insgesamt ist das gewichtete Mittel also eine geeignete Methode für einen großen Parameterbereich.

Games have been investigated as computationally complex problems since the inception of artificial intelligence in the 1950’s. Originally, search-based techniques were applied to create a competent (and sometimes even expert) game player. The search-based techniques, such as game trees, made use of human-defined knowledge to evaluate the current game state and recommend the best move to make next. Recent research has shown that neural networks can be evolved as game state evaluators, thereby removing the human intelligence factor completely. This study builds on the initial research that made use of evolutionary programming to evolve neural networks in the game learning domain. Particle Swarm Optimisation (PSO) is applied inside a coevolutionary training environment to evolve the weights of the neural network. The training technique is applied to both the zero sum and non-zero sum game domains, with specific application to Tic-Tac-Toe, Checkers and the Iterated Prisoners Dilemma (IPD). The influence of the various PSO parameters on playing performance are experimentally examined, and the overall performance of three different neighbourhood information sharing structures compared. A new coevolutionary scoring scheme and particle dispersement operator are defined, inspired by Formula One Grand Prix racing. Finally, the PSO is applied in three novel ways to evolve strategies for the IPD – the first application of its kind in the PSO field. The PSO-based coevolutionary learning technique described and examined in this study shows promise in evolving intelligent evaluators for the aforementioned games, and further study will be conducted to analyse its scalability to larger search spaces and games of varying complexity. / Dissertation (MSc)--University of Pretoria, 2005. / Computer Science / unrestricted