Complex Adaptive Systems of Social Insect Colonies: Emergence of Scaling Social Dynamics Evolution Cooperation (ASUF 30007223)

Project: Research project

Project Details


Complex Adaptive Systems of Social Insect Colonies: Emergence of Scaling Social Dynamics Evolution Cooperation (ASUF 30007223) Complex Adaptive Systems of Social Insect Colonies: Emergence of Scaling, Social Dynamics, and Evolution Cooperation Complex Adaptive Systems of Social Insect Colonies: Emergence of Scaling, Social Dynamics, and Evolution Cooperation Overview Social insects such as ants, bees, wasps and termites, among the most diverse and ecologically important organisms on earth, live in intricately governed societies that rival our own in complexity and internal cohesion1-5. As complex adaptive systems, colonies behave as integrated units and operate as distributed and cooperated systems with no central controller7, such that higher level group organizational patterns are driven in large part by self-organization4-7. Self-organization allows the simple behaviors of individuals to generate complex outcomes for the group with important properties such as resiliency, the ability to recover or maintain function in the face of environmental perturbation, and robustness, the ability to maintain an internal program or trajectory within a dynamic environment1-2,6. These properties have led to an increased interest in the dynamics and organization of social insect colonies not only in biology but also in epidemiology, network routing, optimization theory and robotics8-9. Self-organizational models of behavior have made profound contributions to our understanding of social organization and cooperation4, but the impacts of nonlinear social interaction dynamics on individual and group-level fitness outcomes with respect to the emergence scaling of the colony size are poorly understood and are rarely integrated into models of social behavior and evolution26-29,33-34. Our proposed work will develop novel and new multi-level dynamical network models combined with empirical work to advance our understanding on how emergence of scaling, social dynamics, and evolution cooperation arise, interact, and take effects in social insect colonies as complex adaptive systems. My recent work has been focused on ecological and evolutionary dynamics of social insect colonies including multiscale modeling of the division of labor in social insects27-30,35-39. One of our recent work has discovered that small colonies invest more resources into colony growth, directing more worker ants toward riskier jobs like foraging; In larger colonies, more workers perform safer tasks inside the colony38. We are also developing an evolutionary modeling framework in a dynamical environment to explore how collaborative behavior forms in social groups by comparing two species of ants, one in which queens start colonies on their own and another in which queens work together40. Our current27-30,35-38 and ongoing work40 has laid the groundwork for the proposed research on how social organization scales with size, and how the emergent effects of social interactions (social dynamics) affect social phenotype and fitness, in the evolution of cooperative behavioral strategies. By closely working with the Social Insect Research Group27,30,35,39-40 and the Simon A. Levin Mathematical and Computational Modeling Sciences Center28,38-43 at ASU where collaborations are already taken place, and the ASU-SFI Center for Biosocial Complex Systems launched in 2015, our proposed research program will provide essential contributions to the questions of how group organization scales with size and why animals form cooperative groups which are central in social behavior and evolution. Proposed Research Program Our proposed work will incorporate important components of social insect colonies such as (i). The colony size scaling from small groups of a few individuals to large-scale societies of thousands to millions where the non-linear interaction effects of individual decisions at the local level produce complex colony-wide behaviors including cooperative brood care inside colonies, cooperative nest founding with multiple queens1,4,32,47-48; (ii). The life cycle of colony consisting of the colony founding, ergonomic growth, and colony reproduction stages, where each colony stage is shaped by a distinctive blend of multi-level nonlinear social interactions with potential multi-level fitness effects from individual to colony levels4,32,49-52. More specifically, we will build on a cross-disciplinary collaboration already in place27-30,35-43 and integrate mathematical modeling with experimental biology to address the important and interrelated questions (see Figure 1) on social insect colony as complex adaptive systems: 1. How do task organization and work performance scale with colony size? 2. How does the scaling of work organization affect colony metabolism and growth?3. How do social interaction network structure and the related information flow scale with colony size? 4. What are the emergent dynamics inherent in early social group formation? 5. How are social phenotypes shaped by the transition to cooperative associations? 6. How do costs and benefits of cooperation balance across colony founding and reproduction? We propose to carry out the proposed research over four years. We will develop multiscale dynamical network models in adaptive environments to address the first three questions in the first two years regarding emergence of scaling effects in social dynamics of social insect colonies; then apply our results and extend the evolutionary game theory (EGT) approach to model the phenotypic and fitness outcomes of nonlinear social dynamics, and how these effects influence evolutionary cooperation models to address the last three questions during the remaining two years. 1. Scaling Effects and Social Organization The scaling pattern of metabolic rates scale hypometrically, so that mass-specific metabolic rates decrease with increasing size, has been reported across the wide diversity of animal life, and also for integrated social groups such as social insect colonies; ecological correlates have been reported at scales from populations through ecosystems54-55. Despite their ubiquity, we still lack a fundamental understanding of the behavioral and physiological mechanisms driving these allometries. An emerging theme in biology is that self- organizational processes drive patterns across biological scales4,56-57. Multiple studies show that ant colonies exhibit hypometric scaling of metabolic rates, both in cross-colony comparisons, and as colony size changes during ontogeny65,75. It has been proposed that the lower metabolic rate per gram of social groups reflects economies of scale achieved by larger social groups, either through increased worker polymorphism or via complex social interactions; this benefit has been termed social synergy59. We aim to explore the idea that metabolic scaling with colony size has a mechanistic basis in the collective behavior of their workers and is produced by changes in the organization of work within the colony. The activities of communal and cooperatively breeding taxa are directed simultaneously at individual and group maintenance and reproductive success, generating complexities in the analysis of scaling effects[]. However, workers in a social insect colony, almost like cells within an organism, coordinate in directing most of their activities specifically towards colony function (growth, reproduction, and maintenance)58. The collective activities contributing to colony function are generally categorized as task organization which can essentially be divided into three interconnected but separately measurable characteristics: (a) task allocation, considered as the distribution of the colonys work effort across different tasks (b) division of labor, the degree to which different workers specialize on the different tasks being performed; and (c) task activity level, the duration and intensity of work performance, or total work output, at the individual and/or colony levels. In most social contexts, the relationships between social organization and metabolism are more complex than size or volumetric scaling effects, but it is a challenge to connect scaling changes in social organization to concrete changes in behavior and/or group composition that directly influence metabolism[]. The critical questions center on: What are the changes in colony organization that generate shifts in worker task performance and activity as colony size increases? Are they caused by constraints on internal function, such as limits to resource distribution or alternatively to task-related information flow? Or, do they represent active regulatory changes that contribute to improved colony function? There is currently no model set specifically examining the relationships between colony growth, colony task organization and metabolic scaling. In a model of early colony growth and task organization, Kang et al.27 showed that the division of labor within leafcutter colonies significantly influences the stability of colony growth, especially at low colony sizes. The model provided explanatory power for the observation that colony survival, even in seemingly healthy colonies, is highly erratic at sizes below 100 workers27,67. Models based on metabolic scaling have also examined how ontogenetic changes in worker size and metabolism could produce hypometric changes in colony-level metabolism68,69.. These models illustrate the utility of an energy-based approach, but have not yet integrated scaling changes in colony task organization with predicted metabolic outcomes. Conceptual framework and modeling approaches: Our proposed models, in combination with empirical work will dissect the relationships among colony size, metabolism and the organization of work. By integrating theoretical and empirical analyses, we should be able to better isolate cause and effect among the different interacting components of work organization, and examine their potential impacts on energetic scaling. Our primary approach will be to use a combination of differential equations and related optimizations to explore how components of task organization might generate hypometric metabolic scaling and test whether increased metabolic efficiency can be generated via scaling changes in task organization. In combination, these models will help us determine (1) the principle driving factors in the organization of work across size scales, and (2) how specific component changes in work organization and output affect metabolic scaling. We will also build a set of stochastic spatial explicit models, using the hypothesized relationships between task organization parameters, to assess the effects of individual variation (for example, size and genetic effects), and/or stochastic environment effects. Coupled with our experimental analyses, they provide a powerful methodology for exploring colony organization and metabolic scaling. Significance: The scaling effects of group size on work organization are relevant to all social groups, but are particularly apparent in the social insects. This study will provide an essential contribution to the question of how group organization scales with size, a central question in social behavior1,60-61. Our work will address one of the most open and debated questions in physiological and ecology why the metabolic output of biological systems scales hypometrically with size. 2. Communication Networks Scale with Colony Size to Mediate the Organization of Work The organization of task allocation and division of labor in a social insect colony revolves around its communication networks across different scales61-63. We define topology of network in social insects in three scales: i) large scale-task groups: the communication patterns and strength among different task groups and the emergent new tasks (see the right network of Figure 2); ii) small scale-individual insects: the communication pattern and strength among individual insects and the size of colonies (see the left network of Figure 2); and iii) kinetic scale: as the number of insects grows large, the limiting behavior of the colony. Similarly, we can define node dynamics in three scales: i) large scale: the communication patterns among different task groups and the emergent new tasks; ii) small scale: the decision on which task to perform and the length of its active time; and iii) kinetic scale: the aggregation of individual nodes (ant-agents) engaged in spatially modulated information exchange into task nodes. This interaction network determines the speed of both information and disease flow within the colony53. Both the topology of the network and the role of each worker in the network change over time. The developmental trajectory of each individual worker creates a feedback mechanism that links network structure with agent dynamics. Data show clearly that direct physical contacts (antennations) can generate changes in the number of workers performing a task62,70-72. Almost all work activities, from task switching to recruitment, to assessment of task needs, require workers to interact with others. Therefore, the intricate interaction network among workers provides a unique opportunity for examining the effects of network topology on nodal dynamics and assessing the scaling relationships between colony size and network structure, and particularly the inter-individual interactions within and between tasks. The specific relationships between network attributes and the organization of work as colony size increases is less known. We hypothesize that size-based changes in colony network structure will positively influence the organization of work in three ways: (1) Global network structure and motif substructures scale with colony size to maintain high rates and efficiency of information flow across the colony74. (2) Larger colony networks should increase community structure; i.e. the connectedness among individuals within task subgroups becomes higher relative to their connections outside the group (small world networks66). This would have the effect of increasing modularity of information flow within and between tasks. (3) The physical nest (spatial) structure of colonies contributes positively to network modularity73. To explore how social interaction network structure and the related information flow scale with colony size, we propose to the following modeling approaches combined with experimental work. Multiscale dynamical network modeling approaches: The current experimental work provides baseline information on the possible modular interaction network among individual members and different task groups, which allow us to build network models in different scales to explore the effects of topology of network, the strength and feedback of interaction among task groups as well as essential parameters that are beyond the scope of the experiments. To model network dynamics in different scales, we will use different mathematical approaches. i) Large scale analysis is based on nonlinear differential equations (e.g., Ordinary differential equations, delay differential equations, integral differential equations) which are compartmental models, coupled compartments, viewed as a network of interactions. ii) Small scale analysis is based on the framework of interacting particle systems (individual-based models) which are continuous-time Markov chain models on connected graphs, viewed as a network of interactions, coupled via a set of rules modeling the transition rates at each node (individuals). The study of these models can generate insight on the macroscopic (i.e., division of labor) behavior and spatial patterns that emerge from microscopic (i.e., collective behavior) interactions. and iii) Kinetic model analysis based on aggregating the behavior of the individual agents over space and time into continuous evolution equations for the probability distributions of the various agents in their task groups allows us to connect the characteristic features of the agents to the aggregate features of the task groups. Significance: Social insect colonies have key network attributes such as growth in size, non-random connectivity and emergent properties generated by self-organization2,4-5. Thus, social insects provide not only a powerful system for examining how network dynamics contribute to the evolution of complex biological systems but also a great opportunity to explore how behavior evolves within complex systems2,4-6,9. Social and biological systems must balance network rigidity, which prevents flexibility but promotes robustness, and network dynamics, which promotes flexibility but could lead to unpredictable behavior with no structure. Here we propose multi-scale dynamical network models to investigate how social insect networks achieve this balance closely related to how communication networks scales with colony size to mediate the organization of work across different scales. Our work is a valuable component to explore network structure and social dynamics in social insect colonies and in animal social networks more generally-as they are entering a phase of rapid theoretical and empirical expansion5,63-64. Applications of social networks to infectious disease spread in social insects: Living in societies affects disease transmission, and understanding how infectious diseases in social settings is a crucial area of research for humans directly77. Social insect colonies provide a novel experimental approach to manipulate infection and measure disease transmission. Social colonies have a highly evolved social systems and they are able to effectively control many diseases as are known to optimize the transmission of resources like nectar and protein while reducing pathogen spread77-78. Applying the understanding social network interactions across different scales, we could extend novel dynamic multiscale network models including spatial movements to understand the role of group size, group complexity, and individual contact patterns in driving disease transmission; and therefore explore the important components of social living that promote disease transmission, and those that reduce its spread. These applications are expected to provide useful insights into the mechanisms behind social immunity and disease control in humans and other social species78. 3. Integrating Social Dynamics into Evolutionary Models of Cooperation The question of why animals form cooperative groups is central to our understanding of social evolution16. Theoretical models have focused intensively on the inclusive fitness costs and benefits of sociality10, including game theoretic approaches based on considerations of individual strategy pay-offs12-15, and population genetics models based on considerations of direct and indirect fitness gain15-18. These approaches have made significant contributions to social evolutionary theory. However, such higher-level models can miss the critical effects of social dynamics, which we define as the nonlinear and often complex effects of social interactions on social behavioral phenotypes19-24. Social dynamics play an important role in shaping social behavior across the diversity of cooperative groups, and thus influence social evolution more generally20-22. We propose to explore this idea, by developing a new and more sophisticated modeling approach that captures these interaction effects, with better alignment between experimental biology and its mathematical assumptions. Most models of social evolution begin with simplifying assumptions about the fitness consequences of specific cooperation strategies, and then predict outcomes at the level of societies and/or populations. These assumptions generally avoid explicit consideration of proximate self-organizational processes that amplify social differences and/or generate cascading effects on behavior and social outcomes. However, these dynamics have profound effects on social phenotypes, on patterns of selection, and thus on the evolution of social groups19-23. Social dynamics appear across diverse behavioral contexts, from aggression and dominance, to cooperation and behavioral coordination20,23-26. They can generate extreme costs, such as when aggression escalates into intense conflict20,26. They can also contribute to group coordination in the contexts of task organization and group consensus23-25. The effects of social dynamics change with group size and composition, as well as with life history stage of the social group and its individual group members22-23,25. The proposed research aims to develop a multi-stage evolutionary game theory (EGT) modeling framework linking empirical measures of social dynamics and associated behavioral and fitness outcomes, with its potential role in the evolution of cooperative sociality. Our models will be built based on the empirical system of cooperative nest founding by harvester ant queens for which we can: (1) vary social context by creating mixes of different social phenotypes; (2) directly measure non-linear effects of social context on, and (3) assess the costs and benefits of these dynamics on tokens of individual and group fitness. The harvester ant, Pogonomyrmex californicus, has two contiguous populations in which newly mated queens either primarily found nests alone or form cooperative groups of non-relatives to establish nests21,23,24-26. This taxon allows a rare opportunity to test cooperation models in a naturally occurring population of cooperators and non-cooperators through colony maturity and reproduction, by measuring individual social phenotypes and fitness effects across social contexts. Because queens are nonrelatives25, it also provides a social context not easily captured by traditional kin selection approaches31. Based on the empirical study, we will develop realistic and mechanistic EGT models spanning individual to group levels, and across the different ontogenetic stages of colony life history, to investigate interrelated questions 3-6 about social dynamics in associations of normally cooperative versus single-founding queens, their influence on individual and group selection, and the fitness consequences of cooperation from colony founding through reproduction. Evolutionary game theory (EGT) framework: Game theoretical models essentially consider different social strategies that co-evolve with each other as a mathematical game that has players, strategies, strategy sets, and pay-offs14. In classic EGT models, the players are the individual organisms, strategies are phenotypic or social traits with heritable components, the strategy set is the collection of all evolutionarily feasible strategies for a particular organism, and the pay-off to an individual in groups of nonrelatives is determined by its direct fitness12-14. The evolutionary game deals with the survival of a given strategy within a population of individuals using potentially many different strategies. Our previous work has used this methodology to derive nonlinear equations that describe the population (energy) dynamics of n interacting population (agent) together with the dynamics of (mean) phenotypic traits (or strategies) that serve to characterize all individuals of a particular species and are assumed to have a heritable component28-30. We will extend this EGT approach combined with results from previous two topics to incorporate dynamics with multiple fitness components across the life history stages of: (a) colony founding (cooperation during nest establishment) and (b) reproduction (competition for individual reproductive gain). Our proposed EGT framework includes individuals and/or groups as players; their related social interaction dynamics (e.g. aggression levels); and interacting phenotype dynamics across levels of selection. Significance: The exploration of self-organization and emergence has made profound contributions to our understanding of social organization16-17. However, there has been almost no work examining the role of emergent social dynamics in the early evolution of sociality, or of the interplay of self-organization and selection in shaping social evolution. Our proposed work will integrate these two frameworks, arguing for a paradigm shift in the way we understand the behavioral dynamics and fitness consequences underlying the evolution of cooperation and sociality. It will also provide unique opportunity to explore how well the interactions between cooperative and non-cooperative phenotypes match theoretical models of social evolution during the transition to cooperative sociality. References Cited 1. Anderson C. and McShea D.W. (2001) Individual versus social complexity, with particular reference to ant colonies. Biological Reviews 76:211237. 2. Bonabeau E. (1998). Social insect colonies as complex adaptive systems. Ecosystems 1:437443. 3. Bonabeau E., Theraulaz, G. and Deneubourg, J.L. (1998) Fixed response thresholds and the regulation of division of labor in insect societies. Bulletin of Mathematical Biology 60:753-807. 4. Camazine S, J-L Deneubourg, NR Franks, J Sneyd, G Theraulaz, E Bonabeau (2001) Self-organization in Biological Systems. Princeton University Press, Princeton, NJ. 5. Fewell J.H., 2003. Social Insect Networks, Science, 301, 1867-1870. 6. Fewell J.H., Schmidt, S.K. and Taylor, T. (2009) Division of labor in the context of complexity. In Organization of Insect Societies (eds. Gadau, J. and Fewell, J.H.) Harvard University Press, pp. 483502. 7. Gordon D. M. (2007). Control without hierarchy. Nature 446, 143. 8. Pratt S. C., 2009. Insect societies as model for collective decision making. In: J. Gadau and J.H. Fewell (Eds.), Organization of Insect Societies: From Genome to Sociocomplexity, Harvard University Press, pp. 503-524. 9. Sumpter D.T., 2010. Collective Animal Behavior. Princeton University Press, Princeton, NJ. 10. Hamilton WD (1964) The genetical evolution of social behavior. I, II. Journal of Theoretical Biology 7:1-52. 11. Axelrod R, and WD Hamilton (1981) The evolution of cooperation. Science 211: 1390-1396 12. Maynard Smith J (1982) Evolution and the Theory of Games. Cambridge University Press, Cambridge, UK. 13. Dugatkin L.A, and HK Reeve (1998). Game Theory and Animal Behavior. Oxford University Press. 14. Bijma P, and MJ Wade (2008) The joint effects of kin, multilevel selection and indirect genetic effects on response to genetic selection. Journal of Evolutionary Biology 21:1175-1188. 15. Queller DC (2011) Expanded social fitness and Hamilton's rule for kin, kith, and kind. Proceedings of the National Academy of Sciences 108:10792-10799. 16. West-Eberhard MJ (1979) Sexual selection, social competition, and evolution. Proceedings of the American Philosophical Society 123:222-234. 17. West-Eberhard MJ (2003) Developmental Plasticity and Evolution. Oxford University Press. 18. Moore AJ, ED Brodie III, and JB Wolf (1997) Interacting phenotypes and the evolutionary process: I. Direct and indirect genetic effects of social interactions. Evolution 51:1352-1362. 19. McGlothlin JW, AJ Moore, JB Wolf, and ED Brodie III (2010) Interacting phenotypes and the evolutionary process: III. Social evolution. Evolution 64:25582574. 20. Allen B, MA Nowak and U Dieckmann (2013) Adaptive dynamics with interaction structure. The American Naturalist 181:E139-163. 21. Clark RM, and JH Fewell (2014) Social dynamics drive selection in cooperative associations of ant queens. Behavioral Ecology, art093. 22. Jeanson R, and JH Fewell (2008) Influence of the social context on division of labor in ant foundress associations. Behavioral Ecology 19:567-574. 23. Archer J, and F Huntingford (1994) Game theory models and escalation of animal fights. In The Dynamics of Aggression: Biological and Social Processes in Dyads and Groups (pp. 3-13). L. Erlbaum Hillsdale. 24. Johnson RA (2004) Colony founding by pleometrosis in the semiclaustral seed-harvester ant Pogonomyrmex californicus (Hymenoptera: Formicidae). Animal Behaviour 68:1189-1200. 25. Overson R (2011) Causes and Consequences of Queen Number Variation in the California Harvester Ant Pogonomyrmex californicus. Dissertation, Arizona State University 26. Helms-Cahan S, and JH Fewell (2004) Division of labor and the evolution of task sharing in queen associations of the harvester ant Pogonomyrmex californicus. Behavioral Ecology and Sociobiology 56:9-17. 27. Kang Y, RM Clark, M Makiyama, and JH Fewell (2011) Mathematical modeling on obligate mutualism: leaf-cutter ants and fungus growth during early colony expansion. Journal of Theoretical Biology, 289:116-127. 28. Kang Y, and U Oyita (2014) Dynamics of a single species evolutionary model with Allee effects. Journal of Mathematical Analysis and Applications, 418 (1), 492-515. 29. Kang Y, M Rodriguez, and S Evilsizor (2015) Ecological and evolutionary dynamics of a two-stage social insect model with egg cannibalism. Journal of Mathematical Analysis and Applications, 418 (1), 492-515. 30. Kang Y, and J Fewell (2015) Coevolutionary dynamics of a social parasite-host interaction model: obligatory versus facultative social parasitism. Natural Resource Modeling. In press. 31. Clutton-Brock T (2002). Breeding together: kin selection and mutualism in cooperative vertebrates. Science 296:69-72 32. Cole, BJ (2009) The ecological setting of social evolution: the demography of ant populations In: Gadau J, Fewell JH, eds: Organization of Insect Societies: From Genome to Sociocomplexity. Harvard University Press, Cambridge, MA, pp74-104 33. Moore AJ, ED Brodie III, JB Wolf (1997) Interacting phenotypes and the evolutionary process: I. Direct and indirect genetic effects of social interactions. Evolution 51:1352-1362. 34. Wolf JB, ED Brodie III, AJ Moore (1999) Interacting phenotypes and the evolutionary process: II. Selection resulting from social interactions. American Naturalist 153:254-266 35. Helmkampf M., S. Mikheyev, Y. Kang, J. Fewell and J. Gadau (2016). Gene expression and variation in social aggression in the harvester ant Pogonomyrmex californicus. Submitted to Molecular Ecology. Under Review. 36. Aydogmus O., W. Zhou and Y. Kang (2016). On the preservation of cooperation in two-strategy games with nonlocal interactions. Mathematical Biosciences. Under Revision. 37. Kang Yun, and Guy Theraulaz (2016). Dynamical models of task organization in social insect colonies. Submitted to Bulletin of Mathematical Biology. Minor Revision. 38. Udiani O, N. Pinter-Wollan, and Y. Kang (2015). Identifying robustness in the regulation of foraging of ant colonies using an interaction based model with backward bifurcation. Journal of Theoretical Biology, 365, 61-75. 39. Kang Yun, Krystal Blanco*, Talia Davies**, Ying Wang and Gloria DeGrandi-Hoffman (2016). Disease dynamics of Ho
Effective start/end date10/1/169/30/23


  • James S. McDonnell Foundation (JSMF): $450,000.00


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.