No description available.

In swarm robotics, the control of a group of robots is often fully distributed and does not rely on any leader. In this thesis, we are interested in understanding how to design collective decision making processes in such groups. Our approach consists in taking inspiration from nature, and especially from self organization in social insects, in order to produce effective collective behaviors in robot swarms. We have devised four robotics experiments that allow us to study multiple facets of collective decision making. The problems on which we focus include cooperative transport of objects, robot localization, resource selection, and resource discrimination. We study how information is transferred inside the groups, how collective decisions arise, and through which particular interactions. Important properties of the groups such as scalability, robustness, and adaptivity are also investigated. We show that collective decisions in robot swarms can effectively arise thanks to simple mechanisms of imitation and amplification. We experimentally demonstrate their implementation with direct or indirect information transfer, and with robots that can distinguish the available options partially or not at all.

Many animal societies rely on highly influential keystone individuals for proper functioning. When information quality is important for group success, such keystone individuals have the potential to diminish group performance if they possess inaccurate information. Here, we test whether information quality (accurate or inaccurate) influences collective outcomes when keystone individuals are the first to acquire it. We trained keystone or generic individuals to attack or avoid novel stimuli and implanted these trained individuals within groups of naive colony-mates. We subsequently tracked how quickly groups learned about their environment in situations that matched (accurate information) or mismatched (inaccurate information) the training of the trained individual. We found that colonies with just one accurately informed individual were quicker to learn to attack a novel prey stimulus than colonies with no informed individuals. However, this effect was no more pronounced when the informed individual was a keystone individual. In contrast, keystones with inaccurate information had larger effects than generic individuals with identical information: groups containing keystones with inaccurate information took longer to learn to attack/avoid prey/predator stimuli and gained less weight than groups harbouring generic individuals with identical information. Our results convey that misinformed keystone individuals can become points of vulnerability for their societies.

Cette étude démontre une nouvelle forme de synchronisation des investisseurs sur les marchés des changes qui résulte de deux phénomènes psychologiques : La forme de synchronisation examinée ici est un effet des actions spontanées et décentralisées des individus qui prennent leurs propres décisions basées sur leur compréhension interne du marché qui est, en l'occurrence, une "interprétation commune". Dans ce cas, le comportement des investisseurs n'est pas guidé par de quelconques informations « objectives » concernant le marché ni par leurs interactions sociales intentionnelles qui peuvent naturellement faciliter la synchronisation. Dans la première étude, j'ai étudié l'influence des émotions sur le choix des stratégies d'investissement qui peuvent avoir une incidence sur le schéma dominant de la politique des prix. J'ai découvert qu'il existait un effet d'interaction significative de la valence affective et de la stimulation sur le choix de la stratégie d'investissement. Les émotions positives qui occasionnent un niveau élevé d'excitation peuvent faciliter l'émergence d'une synchronisation dans une tendance haussière. Dans la deuxième étude, j'ai utilisé une combinaison de deux techniques expérimentales : des expériences avec des sujets humains et des simulations par ordinateur, pour étudier la dynamique d'une prise de décision collective dans un modèle simple de marchés financiers. Les expériences montrent à quel point l'interaction de l'historique des prix et de certains mécanismes d'apprentissage peut conduire à l'émergence de préjugés collectifs spontanés. De plus, le fait d'appliquer des simulations par ordinateur sur des données générées par des humains permet d'effectuer une prévision de la synchronisation. Les expériences et le cadre théorique suggèrent de nouvelles voies permettant d'aborder la constitution collective d'un comportement spéculatif. / This work shows a new way of synchronization of investors on exchange markets, which results from two psychological phenomena: 1) emotions and 2) cognitive mechanisms. The way of synchronization considered here is an effect of spontaneous and decentralized actions of individuals, who make their own decisions based on their internal understanding of the market which happens to be “shared interpretation”. In such case, the behavior of investors is not guided by any kind of 'objective' market information or their intentional social interactions, which may naturally facilitate synchronization. In the first study I investigated the influence of emotions on choice of investment strategies that may impact a dominant pattern of the price behavior. I found that there is a significant interaction effect of affective valence and arousal on the selection of investment strategy. Positive emotions, which cause high arousal, may facilitate emergence of synchronization in the direction of uptrend. In the second study, I used a combination of two experimental techniques: experiments with human subjects and computer simulations, to study the dynamics of collective decision-making in a simple financial markets' model. The experiments show how interplay of certain price history and learning mechanisms can lead to the emergence of spontaneous collective biases. Additionally, applying computer simulations to the data generated by the humans enables prediction of synchronization. The experiments and a theoretical framework suggest new ways to access the pathways involved in a collective formation of speculative behavior.

This thesis focuses on the phenomenon of strategic manipulation in collective decision making, i.e. situations where it is beneficial for certain members of the group to derive from their sincere preferences in the voting process. This manipulation complicates the process of collective decision making, leading to situations where lying becomes a successful strategy that some mambers of the group can use to gain an advantage over others. In the thesis, examples of strategic manipulation and a summary of up to date research in the field of social choice theory are included. The main objective is to create a simulation model that can be used to quantify the probability of a successful manipulation. A Monte Carlo simulation using this model is performed on ten voting procedures. Analysed procedures include the most common and intuitive ones, more sophisticated systems created by experts of social choice theory and the most modern attempts to improve democracy by Karel Janeček.

abstract: An insect society needs to share information about important resources in order to collectively exploit them. This task poses a dilemma if the colony must consider multiple resource types, such as food and nest sites. How does it allocate workers appropriately to each resource, and how does it adapt its recruitment communication to the specific needs of each resource type? In this dissertation, I investigate these questions in the ant Temnothorax rugatulus. In Chapter 1, I summarize relevant past work on food and nest recruitment. Then I describe T. rugatulus and its recruitment behavior, tandem running, and I explain its suitability for these questions. In Chapter 2, I investigate whether food and nest recruiters behave differently. I report two novel behaviors used by recruiters during their interaction with nestmates. Food recruiters perform these behaviors more often than nest recruiters, suggesting that they convey information about target type. In Chapter 3, I investigate whether colonies respond to a tradeoff between foraging and emigration by allocating their workforce adaptively. I describe how colonies responded when I posed a tradeoff by manipulating colony need for food and shelter and presenting both resources simultaneously. Recruitment and visitation to each target partially matched the predictions of the tradeoff hypothesis. In Chapter 4, I address the tuned error hypothesis, which states that the error rate in recruitment is adaptively tuned to the patch area of the target. Food tandem leaders lost followers at a higher rate than nest tandem leaders. This supports the tuned error hypothesis, because food targets generally have larger patch areas than nest targets with small entrances. This work shows that animal groups face tradeoffs as individual animals do. It also suggests that colonies spatially allocate their workforce according to resource type. Investigating recruitment for multiple resource types gives a better understanding of exploitation of each resource type, how colonies make collective decisions under conflicting goals, as well as how colonies manage the exploitation of multiple types of resources differently. This has implications for managing the health of economically important social insects such as honeybees or invasive fire ants. / Dissertation/Thesis / Doctoral Dissertation Biology 2019

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第19920号 / 理博第4220号 / 新制||理||1606(附属図書館) / 33006 / 京都大学大学院理学研究科生物科学専攻 / (主査)教授 古市 剛史, 教授 湯本 貴和, 教授 平井 啓久 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

La prise de décision collective est un processus dans lequel un groupe d'individus, ayant des intérêts différents, se réunit pour trouver une solution collective à un problème. Ce processus est inhérent aux activités de toute organisation politique, économique ou sociale. Le développement de l'Intelligence Artificielle notamment les Systèmes Multi-Agents a permis la modélisation et l'automatisation des processus de prise de décision afin de mieux comprendre et d'analyser leur fonctionnement. Une décision collective peut être prise par vote ou par négociation. Dans le cadre de cette thèse, nous abordons les mécanismes de négociation multilatérale pour la prise de décision collective basés sur des approches heuristiques. Les agents construisent la solution à leur problème à travers leurs interactions à la différence des modèles basés sur la théorie des jeux dont l'espace des solutions est supposé connu par tous les agents. Le problème des négociations heuristiques réside dans les mécanismes de raisonnement des agents dont la complexité augmente lorsque le nombre d'agents et d'attributs à négocier devient important. L'objectif de cette thèse est ainsi de proposer des mécanismes de négociation décentralisés (sans médiateur) et distribués en mettant en exergue l'aspect organisationnel des agents. Notre approche s'inspire du concept diviser pour régner et permet aux agents de négocier de façon incrémentale. Le but est de faciliter la recherche d'accords et de limiter la complexité du raisonnement des agents. Les travaux de cette thèse ont abouti à trois contributions abordant la négociation multi-agents sous différents angles tels que l'organisation des agents, le protocole d'interaction et les stratégies de concession et de choix de solutions équitables et justes. Pour valider nos propositions, nous avons implémenté sous JavaJade les mécanismes de négociation proposés. Les critères de performance que nous avons évalués sont, notamment, la convergence, le temps de négociation et la qualité de la solution. Nous avons comparé nos modèles avec ceux existants et les résultats obtenus montrent leur efficacité pour l'obtention des accords entre les agents / Collective decision making is a process in which many participants with different interests interact in order to build a solution to their problem. It is inherent to many organisations and companies. Nowadays, the advances in Artificial Intelligence, notably, Multi-Agents Systems enabled the automation of decision-making processes in order to analyse and to better understand how these mechanisms work. A collective decision may be made by using a voting system or by using negotiation. In this thesis, we focus on multilateral negotiation for collective decision making by proposing negotiation models. The proposed models based on heuristic approach. The agents interact with them in order to build a solution to their problem. This context is different from models based on game theory where the set of possible solutions are supposed to be known by all agents. So heuristic negotiation issue is that agents' reasoning may be very complex. This complexity grows where the number of agents and issues to be negotiated are important. The goal of this research work consists of devising negotiation mechanisms where agents'interaction are fully decentralized. We focus on organisation aspect of the multi-agent system by using divide and conquer approach in order to reduce the negotiation complexity and hence to facilitate research of agreements. Our works tackle negotiation under different contexts which lead us to bring three contributions which focus on agents' organization, interaction protocols, negotiation object, concession strategies and effective and fair solution concept. The proposed mechanisms are implemented in JavaJade. We analyse the convergence of the negotiation, negotiation time and quality of the solution. Our models are compared with a centralized approach where all of the agents are gathered around one group to negotiate. Our empirical analyses show that our propositions allow the agents to reach collectives agreements

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-07-28T20:13:54Z No. of bitstreams: 1 nilhiangoncalvesdealmeida.pdf: 901056 bytes, checksum: f81be41f6ec01326237e68fcf570b703 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-07-29T11:42:09Z (GMT) No. of bitstreams: 1 nilhiangoncalvesdealmeida.pdf: 901056 bytes, checksum: f81be41f6ec01326237e68fcf570b703 (MD5) / Made available in DSpace on 2016-07-29T11:42:09Z (GMT). No. of bitstreams: 1 nilhiangoncalvesdealmeida.pdf: 901056 bytes, checksum: f81be41f6ec01326237e68fcf570b703 (MD5) Previous issue date: 2016-02-25 / CNPq - Conselho Nacional de Desenvolvimento Científico e Tecnológico / Avaliação do processo de tomada de decisão na determinação de rotas em Acromyrmex subterraneus (Forel, 1893) (Hymenoptera, Formicidae). O forrageamento é imprescindível à sobrevivência e reprodução dos animais, devendo ser entendido como um imperativo biológico. Operárias de formigas cortadeiras utilizam de forma sinérgica informações individual (memória) e social (feromônio de trilha) a fim de selecionar e transportar o alimento. Este transporte ocorre por meio de uma rede de trilhas químicas e físicas, estas por sua vez apresentam várias bifurcações. Em cada bifurcação a tomada de decisão sobre qual rota seguir pode estar relacionada com (i) o comprimento de cada ramo da trilha, (ii) a qualidade do alimento ao final de cada ramo e (iii) a hierarquia estabelecida na utilização das informações disponíveis. Investigou-se a influência destes três fatores na determinação de rotas de forrageamento utilizando sistemas de trilhas artificiais. Verificou-se a escolha de rotas em três pontos de bifurcação consecutivos que conduziam as forrageiras ao longo de ramos curtos ou longos ao recurso. A seleção do ramo curto ocorreu nas três bifurcações, tanto nas fases de recrutamento das escoteiras, na qual não há transporte de carga foliar, quanto na das forrageiras, quando ocorre o transporte. No entanto, na bifurcação mediana a proporção de operárias no ramo curto foi menor em relação às bifurcações das extremidades. Com relação ao efeito da qualidade do alimento, avaliou-se a tomada de decisão das forrageiras em dois pontos de bifurcação consecutivos dos quais divergiam rotas curtas e longas e havia acesso a recursos com duas concentrações de sacarose. Alternando-se a disposição dos recursos nos ramos curto e longo, foi verificado o transporte preferencial do recurso de maior qualidade independente se este estava no ramo curto ou longo, evidenciando que a qualidade energética do recurso alimentar foi determinante para que ocorresse seu transporte. Para verificar a ocorrência de uma ordem hierárquica na utilização das informações individual e social, foi utilizada uma trilha em Y na qual apenas um ramo conduzia ao recurso alimentar. Este ramo teve uma seção recoberta por papel filtro onde houve aplicação de feromônio de trilha pelas operárias, estimada pela contagem do fluxo. Após uma forrageira individualmente marcada (operária-alvo) realizar 1, 3 ou 5 visitas ao recurso, o papel filtro foi transferido para o ramo que não conduzia ao alimento. Registrou-se o tempo gasto pela operária-alvo na seleção de ramo e a frequência de escolhas corretas (ramo com alimento). Independente do número de visitas prévias da operária-alvo ao recurso, a frequência de escolhas certas foi alta. Também o número de visitas não influenciou no tempo para a tomada de decisão. No entanto, quanto maior foi o fluxo de operárias, o qual reflete a concentração de feromônio, maior foi o tempo para a tomada de decisão. Assim, ao conflitar as informações individual e social, verificou-se que a informação individual foi priorizada. O presente estudo demonstra que rotas curtas são preferencialmente utilizadas, mas a qualidade do alimento altera a seleção destas rotas. Ademais, a informação individual sobre a presença do alimento foi determinante na seleção de rotas, sendo priorizada em relação à informação social. / (Evaluation of the decision making process of determining route in Acromyrmex subterraneus (Forel, 1893) (Hymenoptera, Formicidae). Foraging is essential to animal survival and reproduction, and must be understood as a biological imperative. Leaf-cutting ant workers mainly use individual (memory) and social (trail pheromone) information in a synergic way in order to select and transport food, as well as travelling along trails which have many bifurcations. At each bifurcation, the decision making about which route to follow is related to (i) the length of each trail branch, (ii) the food quality at the end of each branch and (iii) the established hierarchy on using available information. The influence of these three factors on foraging route determination was investigated using artificial trail systems. Route selection was verified in artificial trails with three consecutive bifurcations, which led the foragers to the resource through short or long branches. Selection of the short branch occurred at the three bifurcations in both the scout recruitment and the forager recruitment phase. However, at the central bifurcation the proportion of workers at the short branch was lower in relation to the end bifurcations. With regards to food quality effect, forager decision making was evaluated at two consecutive bifurcations from which diverge short and long routes and there was access to resources with two saccharose concentrations. Shifting the resource disposition at short and long branches, preferential transport of high quality food was verified irrespective of whether it was at the short or the long branch, showing that energetic quality of food resource was determinant to its transport occurrence. To verify if there was a hierarchy order on using the individual and social information, a Y-trail was utilized where only one branch led to the food resource. This branch has a section covered by filter paper where there was trail pheromone deposition by workers, estimated by worker flow counting. After an individually marked forager (target-worker) makes 1, 3 or 5 trips to the resource, filter paper was transferred to the branch which did not lead to the food. The time expended by the target-worker for branch selection and right choice (branch with food) frequency was registered. Regardless of the target-worker’s previous trips to the resource, right choice frequency was high. Also, the previous trip number did not influence the time expended for the decision making. However, the higher the worker flow, the more time expended on decision making. Thus, by conflicting individual and social information, it has been verified that the individual one was prioritized. This study demonstrated that short routes are preferentially selected, but food quality alters route selection. Also, individual information about food presence was crucial on route selection, being prioritized in relation to social information.

In this thesis, we used the American cockroach study the relationship between animal personality and collective behaviour. These questions do not only regard domiciliary cockroaches or insects; in fact, the questions we discuss here have a large scope and concern most of the known gregarious species distributed in several taxa. One of our aims is to develop general tools, methods and analyses that could be used for the study of collective behaviour.We show evidence of personality in several behavioural traits and different contexts in the American cockroach. These behavioural differences were observed in a context without social interactions during the daytime (isolated condition) and nighttime (solitary exploration behaviour) and in a social context (rate of joining a shelter and sheltering time). Based on the short-term experiments, this personality can be understood as differences in the probability of joining a shelter. We show that the behavioural variability existing in a population of domiciliary cockroaches is very high. The composition of personalities within a group can lead to group personality – consistent differences in group behaviour. For instance, groups composed of different behavioural profiles show different collective dynamics. Regarding the long-term experiments, our results show that the individuals that were aggregated at a site that was repeatedly disturbed by a lighting stimulus during their resting period showed slow migration to a new shelter, which allowed the initial aggregation site to remain the site of choice for a few days. Moreover, the disturbance regime did not influence the group's global activity rhythm. At the individual level, we observed interindividual differences (personalities) in terms of their position prior to the disturbance but not for the different steps of the fleeing behaviour itself. In addition, we show that thigmotaxis affects the reaction time to the disturbance: individuals near the walls of the shelter react more slowly thanindividuals in the centre. Finally, an approach coupling modelling and experimental data shows that behavioural variability plays a secondary role during migration dynamics, thus highlighting the plasticity of personality traits depending on the context. / Un des comportements collectifs les plus répandus, qu’il s’agisse de vertébrés (mammifères, oiseaux, poissons), d’insectes ou encore de bactéries, est la tendance des individus à se regrouper. Les causes proximales de ces rassemblements ou agrégats, c'est-à-dire les mécanismes à leur origine, ont retenu moins d’attention que l’étude des causes ultimes. De plus, dans la plupart des études portant sur les processus d’agrégation, notamment chez notre organisme modèle, la blatte, la diversité des comportements individuels est souvent sous-estimée ou ignorée. L’objectif de cette thèse est l’étude du rôle des différences interindividuelles, également connues comme personnalité animale ou syndromes comportementales, dans les processus de prise de décision collective et notamment la formation des agrégats.Pour cela, nous avons utilisé des groupes de mâles de la blatte américaine Periplaneta americana. Ces insectes peuvent s’agréger dans un ou plusieurs abris et présenter des différences dans leur réponse au milieu, aux conspécifiques et dans le partage de l’information. Dans le cadre de prises de décisions collectives, nous avons mis en évidence une personnalité individuelle mais également au niveau du groupe. Celles-ci se manifestent dans la recherche d’un abri dont certains individus jouent un rôle clef. Nous avons montré que les groupes présentent une stabilité qui est observable au niveau de la dynamique d'agrégation et qui dépend de la distribution des personnalités au sein de ceux-ci. De plus, nous remarquons que ces différences de personnalités au sein des groupes affectent la prise de décision collective, notamment la vitesse du choix et le nombre total d’individus abrités. De manière surprenante, la composition des groupes n’affecte pas la probabilité que ceux-ci atteignent un consensus: la majorité des blattes étant agrégées sous le même abri.Concernant les influences croisées entre l’effet sociale et la personnalité des blattes, nous montrons que les effets sociaux tendent à supprimer les différences inter-individuelles et créent une corrélation entre le comportement de chaque individu et le comportement du groupe. De plus, nous observons que la présence de différentes personnalités dans un groupe augmente les amplifications sociales, celles-ci étant dues aux interactions entre les membres du groupe. Finalement, nous abordons la question du rôle de la personnalité sur le comportement de fuite lors des perturbations lumineuses et sur la dynamique d’émigration quand le site de repos est régulièrement perturbé. Nous avons montré l’existence de personnalités exprimées pendant la phase active nocturne et de différences interindividuelles au niveau du thigmotactisme pendant la journée (phase passive de repos et d’agrégation). Curieusement, ces différences inter-individuelles ne sont pas observées lors du comportement de fuite et au niveau de la dynamique global d’émigration. Dans la dernière partie de notre thèse, nous discutons, en particulier, des synergies et des conflits entre les différentes personnalités et les dynamiques collectives et avançons l’hypothèse que les phénomènes que nous avons mis en évidence sont partagés par de nombreuses espèces grégaires. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished