Syftet med föreliggande studie är att undersöka Twitch, en plattform för live-strömmat dataspelande, i förhållande till dess aktörers interaktioner. För att avgränsa studien har ett av de hundratals spel som strömmas på Twitch valts ut; Starcraft II. Genom att behandla Starcraft II & Twitch som ett spänningsfält där polariteter, kapital och dess aktörer strukturerar fältet, utformas en teoretisk modell utifrån Bourdieus fältteori (1984; 1986; 1990) och Goffmans (1959) teorier för mikrosocialt handlande. Det framkommer att fältet går att analysera som ett kulturellt sådant, där professionella gamers och personligheter samt cyniska och oförställda framföranden står i motsatsförhållande. Pro-gamers är fältets rena kulturella kapital medan personligheter korrumperar sitt kapital med socialt och ekonomiskt kapital för att bli erkända på fältet. Cyniska framföranden tar avstånd från ’rollen’ som strömmare medan oförställda framföranden omfamnar den. Fältets viktiga interaktionspunkter och markörer mellan strömmare och tittare är: teammaskopier, moderatorstatus, spelmässigt anknutna symboler för skicklighet, spelregion, emoticons, memes, exkludering och inkludering, Tittarna investerar kapital (tidsmässigt, ekonomiskt, socialt och kulturellt) för att bli del av de olika gemenskaperna som existerar dels på Twitch, men också överlappar på plattformar relaterade till Starcraft II. Sammanfattningsvis struktureras fältet utefter dels spelmässig skicklighet och kunnighet och dels distansen till rollen som strömmare.

Artificiell intelligens är ett område inom datavetenskap som försöker skapa intelligenta system eller system som simulerar intelligens. Sådana system är intressanta för konsumenter eftersom de kan utföra uppgifter som annars krävt mänsklig inblandning.Spelindustrin hjälper till att driva utvecklingen av AI framåt när spelare fortsätter att förvänta sig mer engagerande och verklighetstrogna upplevelser.Akademiskt är spel användbara för studier av artificiell intelligens på grund av att de är relativt simpla. Men även om spel är enklare än verkligheten är det fortfarande en svår uppgift att skapa en artificiell intelligens som kan matcha en mänsklig motståndare.En populär genre av spel är strategispel, exempel på dessa är Age of Empires och Starcraft. I denna rapport undersöks en annorlunda ansats för att lösa problemet med resurshantering för denna typ av spel med hjälp av ett artificiellt neuralt nätverk som klassificerar spelets tillstånd. Detta har inte utforskats tidigare och målet är att ta reda på hurvida det är möjligt. För att träna nätverket används backpropagation i samband med Q- learning vilket gör inlärningen unsupervised. Så författarnas frågeställning är följande: Kan ett Q- nätverk användas för att hantera resursallokeringen för en bot i Starcraft Broodwar?För att kunna se hur väl Q- nätverket löser problemet utförs ett experiment med två olika botar där den ena spelar starcraft med samma möjligheter som en spelare och den andra en förenklad version. Experimentet går ut på att samla data från botarnas träning för att se om de förbättras eller inte. Som kontroll används två extra botar som slumpmässigt väljer handlingar.Resultatet av experimentet var flera grafer som visade botarnas prestanda på olika sätt och hur många spel de vunnit och sannolikheten för vinst. Med stöd av resultatet är det inte möjligt att se någon verklig förbättring i botarnas spelande med 0.7% respektive 0.4% chans för vinst mot standard AI:n. Resultatet visar dessutom att en av botarna är mycket sämre än en som slumpat fram handlingar.Dessutom visade det sig att träningen tog alldeles för lång tid. Om experimentet får mer tid kanske det skulle visat att tekniken är möjlig men först efter orimligt lång tid vilket skulle göra den oanvändbar i praktiken. Om detta hade lyckas hade det inneburit att man skulle kunna skapa bättre AI för strategispel som anpassar sig efter spelaren och kan generalisera när den ställs inför en situation som inte var planerad av utvecklarna.Men i denna studie förblev botarnas beteende mer eller mindre stokastiskt så svaret på frågeställningen är att det inte är möjligt. / Artificial intelligence is a field of computer science that tries to create intelligent systems or systems that simulate intelligence. Such systems are attractive to consumers because they can perform tasks that would otherwise have required human intervention.The gaming industry is helping to drive the development of AI forward as players continue to expect more immersive and lifelike experiences.Academically games are useful for the study of artificial intelligence because they are relatively simple. But even if the game is simpler than the reality, it is still a difficult task to create an artificial intelligence that can match a human opponent.A popular genre of games is strategy, examples of which are Age of Empire and StarCraft. This report examines a different approach to solve the problem of resource management for this type of game with the help of an artificial neural network to classify the game state. This has not been explored previously, and the goal is to find out whether this approach is feasible or not. To train the networks back-propagation is used in conjunction with Q-learning which makes learning unsupervised. So the authors’ research-question is: Can a Q network be used to manage the resource allocation for a cure in StarCraft Broodwar?To see how well the Q networks solve the problem an experiment was conducted with two different bots where one play StarCraft with the same opportunities a player would have and the other a simplified version. The experiment consists of collecting data from the bots training to see if they improve or not. As a control, two additional bots are used with a completely random policy.The results of the experiment were several graphs showing the bots performance in different ways but most importantly, the number of games won and the probability of winning. With the support of the result, it is not possible to see any real improvement in bot gameplay with 0.7% and 0.4% chance to win against the default AI. The results also show that one of the neural net bots performed much worse than the one with random actions.Moreover, the training turned out to be far too long. If the experiment had more time maybe it would have shown that the technology is possible, but still, only after an unreasonably long time, which would make it useless in practice. If this had been successful it would have meant that we might create better AI for the strategy games that adapts to the player and can generalize when faced with a situation that was not planned by the developers.But in this study the bots behavior remained more or less stochastic so the answer to the research- question is that it is not possible.

This paper is about the possibility to use evolution to make a StarCraft AI better in some areas by using genetic programming. We aimed to use genetic programming to evolve the numbers of squad units, bunkers and turrets, which are an important part of a successful StarCraft AI. We have built a separate application for handling the evolution. This application runs in parallel with StarCraft and modifies files based on the data recieved from a played game. This is good for safety, since if StarCraft crashes the evolution is just stalled not lost. Our tests ran over the course of a few weeks. A combination of a relatively small amount of time, for something very time-consuming, and a lack of experience with genetic programming resulted in a small amount of results. The conclusion is that it is possible to improve an StarCraft AI with genetic programming, however it takes a lot of time. / Denna uppsats handlar om möjligheten att använda evolution att göra en StarCraft AI bättre i vissa områden med hjälp av genetisk programmering. Vi siktade på att använda genetisk programmering att utveckla antalet trupp enheter, bunkrar och torn, som är en viktig del av en framgångsrik StarCraft AI.

Context. Interest in and research on neural networks and their capacity for finding solutions to nonlinear problems has increased greatly in recent years. Objectives. This thesis attempts to compare competitive coevolution to traditional neuroevolution in the game StarCraft: Brood War. Methods. Implementing and evolving AI-controlled players for the game StarCraft and evaluating their performance. Results. Fitness values and win rates against the default StarCraft AI and between the networks were gathered. Conclusions. The neural networks failed to improve under the given circumstances. The best networks performed on par with the default StarCraft AI.

Context. Games are becoming larger and the amount of assets required is increasing. Game studios turn toward procedural generation to ease the load of asset creation. After the game is released the studios want to extend the longevity of their creation. One way of doing this is to open up the game for community created add-ons and assets or utilize some procedural content generation. Both community created assets and procedural generation comes with a classification problem to filter out the undesirable content. Objectives. This thesis will attempt to create a method to evaluate community-generated StarCraft maps with the help of machine learning. Methods. Manually extracted metrics from StarCraft maps and ratings from community repositories. This data is used to train neural networks using NeuroEvolution of Augmenting Topologies (NEAT). The method will be compared with Sequential Minimal Optimization (SMO) and ZeroR. Results and Conclusions. The problem turned out to be more difficult than initially thought. The results using NEAT are marginally better than SMO and ZeroR. The suspected reason for this is insufficient input data and/or bad input parameters. Further experimentation could be conducted with deep learning to try to find a suitable solution for this problem.

This thesis investigates the performance of the option critic (OC) framework combined with behaviour trees (BTs) in Starcraft 2 mini-games when training time is constrained by a time frame limited by consumer level hardware. We test two such combination models: BTs as macro actions (OCBT) and BTs as options (OCBToptions) and measure the relative performance to the plain OC model through an ablation study. The tests were conducted in two of the mini-games called build marines (BM) and defeat zerglings and banelings (DZAB) and a set of metrics were collected, including game score. We find that BTs improve the performance in the BM mini-game using both OCBT and OCBToptions, but in DZAB the models performed equally. Additionally, results indicate that the improvement in BM scores does not stem solely from the complexity of the BTs but from the OC model learning to use the BTs effectively and learning beneficial options in relation to the BT options. Thus, it is concluded that BTs can improve performance when training time is limited by consumer level hardware. / Denna avhandling undersöker hur kombinationen av option critic (OC) ramverket och beteendeträd (BT) förbättrar resultatet i Starcraft 2 minispel när träningstiden är begränsad av konsumenthårdvara. Vi testar två kombinationsmodeller: BT som makrohandlingar (OCBT) och BT som options (OCBToptions) och mäter den relativa förbättringen jämte OC modellen med en ablationsstudie. Testen utfördes i två minispel build marines (BM) och defeat zerglings and banelings (DZAB) och olika typer av data insamlades, bland annat spelpoängen. Vi fann att BT förbättrade resultatet i BM på båda hierarkiska nivåerna men i DZAB var resultaten ungefär lika mellan de olika modellerna. Resultaten indikerar också att förbättringen i BM inte beror bara på BT komplexitet utan på att OC modellen lär sig att använda BT och lär sig options som kompletterar dess BT options. Vi finner därför att BT kan förbättra resultaten när träningen är begränsad av konsumenthårdvara.

Intelligent robots are rational agents. The rationality of robots working cooperatively is significantly different from robots working independently. Cooperation between intelligent robots requires the high level of reasoning and complex interactions for successful operations. The required reasoning process includes knowledge representation and sharing as well as the ability to understand the context of a situation. The reasoning process heavily influences on the planning of deciding what actions need to be taken. Goal deliberation and planning is the process that deals with those requirements. This dissertation investigates the problem of goal deliberation and planning to enable such cooperation between goal-oriented intelligent robots, working as a team. The dissertation then proposes a multi-robot system model that embraces results of the investigation. The proposed model is realized on the top of the platform ‘robot operating system’ (ROS). The implemented system, named ‘goal-oriented multi agent systems’ (GOMAS), is demonstrated with the computer game, StarCraft II. Units in StarCraft II are individually controlled by the GOMAS robots and work cooperatively to attain a set of goals given from operators. The demonstration with the three different scenarios validates that the GOMAS system successfully and efficiently deliberates and plans the given goals.

Clausewitz and Jomini are two big figures in the military science community. Their principles of the concentration of force are still prevalent today in the further development of principles. With this in mind there have been some studies where computer games have been used to improve military training. The study aims to study how the principles of war can lead to victory in the real-time strategy game Starcraft II. The purpose of this study to contribute to the body of scientific knowledge with using computer games to increase the understanding of the principles of war. The method used in this study is a quantitative content analysis to gather data for analysis in the SPSS-program. The results of the study were that all the use of principles that were chosen, with the exception of surprise, correlated with victory. It also showed that spatial ability leads to higher win probability, this find lowers the reliability of the study since it cannot prove to what extent this leads to victory. However, this result is not applicable in the physical world, it can only contribute to a theoretical understanding of the principles of war.

Communication in team games between human player is important, but has been disregarded in teammate bots for Real-Time Strategy (RTS) games. Control over the team-mate bots have existed for a while in other genres than RTS and if implemented correctly it adds another dimension to the game. In this study we investigate whether players think it is more fun to play with an RTS bot that communicates its intentions and reasons and is controllable by the human player. But also what features are liked, disliked, missed to provide guidelines to future researchers and companies. For this we create a StarCraft RTS bot with communication and control abilities. The experiment consists of four scenarios, turning on/o communication and control, to conclude what players think is fun. All testers agreed on communication being important and more fun to play with. Beginners did, however, not like the control feature as they already had enough on their mind whereas experienced players preferred having some control over the bot We conclude that communication is an important role in team games, including RTS games. More work needs to be done how to integrate control so that beginners do not feel overwhelmed and at the same time experienced players do not feel restrained by too simple control commands.

Developing real-time strategy game AI is a challenging task due to that an AI-player has to deal with many different decisions and actions in an ever changing complex game world. Humans have little problem when it comes to dealing with the complexity of the game genre while it is a difficult obstacle to overcome for the computer. Adapting to the opponents strategy is one of many things that players typically have to do during the course of a game in the real-time strategy genre. This report presents a finite-state machine based solution to the mentioned problem and implements it with the help of the existing Starcraft: Broodwar AI Opprimobot. The extension is experimentally compared to the original implementation of Opprimobot. The comparison shows that both manages to achieve approximately the same win ratio against the built-in AI of Starcraft: Broodwar, but the modified version provides away to model more complex strategies.