Insights into the Impact of Social Networks on Evolutionary Games

In this chapter, we explore the use of evolutionary game theory (EGT) (Weibull, 1995; Taylor & Jonker, 1978; Nowak & May, 1993) to model the dynamics of adaptive opponent strategies for large population of players. In particular, we explore effects of information propagation through social networks in Evolutionary Games. The key underlying phenomenon that the information diffusion aims to capture is that reasoning about the experiences of acquaintances can dramatically impact the dynamics of a society. We present experimental results from agent-based simulations that show the impact of diffusion through social networks on the player strategies of an evolutionary game and the sensitivity of the dynamics to features of the social network.

Cellular Automata

In this chapter we present a view of cellular automata (CAs) as the quintessential complex system and how they can be used for complex systems modelling. First we consider theoretical issues of the complexity of their behaviour, discussing the Wolfram Classification, the Langton, lambda parameter and the edge of chaos. Then we consider the input entropy as a way of quantifying complex rules. Finally we contrast explicit CA modelling of geophysical systems with heuristic particle based methods for the visualisation of lava flows.

Agent Strategies in Economy Market

IThis chapter presents a method on modelling the economy market using agent-based representation and iterated prisoner’s dilemma (IPD). While IPD has been used widely in various economic problems, most of the studies were based on quantitative data which could be deductive and inappropriate. The main objective of this chapter is to present a unique agent based approach which places lower demand on data using IPD to model the complexity of the economy market. We create a simulated market environment with agents acting as firms to perform transactions among each other with chosen IPD strategy. From empirical results, we investigate strategic interactions among different firms. In the concluding remarks, we present our observations on the qualities of a winning strategy.

FinSim

We present in this chapter an overview of a financial system model (FinSim) created by the authors at the Los Alamos National Laboratory. The purpose of this model is to understand the impacts of external disruptions to the financial system, in particular disruptions to telecommunication networks and electric power systems; and to model how those impacts are affected by the interactions between different components of the financial system, e.g. markets and payment systems, and by individual agents actions and regulatory interventions. We use agent-based modeling to represent the interactions within the financial system and the decision-making processes of banks and traders. We model explicitly message-passing necessary for execution of financial transactions, which allows a realistic representation of the financial system dependency on telecommunications. We describe implementation of the payment system, securities market and liquidity market components; and present a sample telecommunications disruption scenario and its preliminary results.

Building Complex Adaptive Systems

Agent Oriented Software-Engineering (AOSE) proposes the design of distributed software systems as collections of autonomous and pro-active actors, so-called agents. Since software applications results from agent interplay in Multi-Agent Systems (MAS), this design approach facilitates the construction of software applications that exhibit self-organizing and emergent dynamics. In this chapter, we examine the relation between self-organizing MAS and Complex Adaptive Systems (CAS), highlighting the resulting challenges for engineering approaches. We argue that AOSE developers need to be aware of the possible causes of complex system dynamics, which result from underlying feedback loops. In this respect current approaches to develop SO-MAS are analyzed, leading to a novel classification scheme of typically applied computational techniques. To relieve development efforts and bridge the gap between top-down engineering and bottom-up emerging phenomena, we discuss how multi-level analysis, so-called mesoscopic modeling, can be used to comprehend MAS dynamics and guide agent design, respectively iterative redesign.

The Intelligence of Rumors

This chapter presents a theory of reputation seen as the result of evaluation spreading in a multi-agent system (MAS). In particular the capacities of agents that spread reputation have been analyzed and decomposed in their atomic parts (pragmatic, memetic, epistemic). For each decision we state and verify several claims, using the methodologies that allowed best to capture the relevant aspects of the theoretical statements. Our major claim is that only when considering agents’ architectures and roles it is possible to find out which regulatory patterns cannot emerge. Reputation is argued to be no exception to this rule. Therefore, in order to interact with each other, agents need evaluation concerning which partner to choose. We will describe how by using various kinds of methodologies that account for theory fragments a coherent picture can be observed and how interdisciplinary can help to account for complex intelligent phenomena among adaptive social systems.

Unique Applications of Multi-Agent Models in Uncovering Language Learning Processes

Multi-scale “artificial societies” are constructed to examine competing first- and second-language acquisition-based theories of creole language emergence. Sociohistorical conditions and psycholinguistic capacities are integrated into the model as agents (slaves and slave-owners) interact. Linguistic transmissions are tracked, and grammar constructions are charted. The study demonstrates how a CAS approach offers clear indications for computational solutions to questions of language change and formation.

Innovation Diffusion Among Heterogeneous Agents

In this paper we apply Agent-Based Modelling (ABM) to capture the complexity of the diffusion process depicted in Medical Innovation, the classic study on diffusion of a new drug Tetracycline by (Coleman, Katz, & Menzel, 1966). Based on our previous model with homogenous social agents, Gammanym (Ratna, Dray, Perez, Grafton, Newth, & Kompas, 2007), in this paper we further our analysis with heterogenous social agents who vary in terms of their degree of predisposition to knowledge. We also explore the impact of stage-dependent degrees of external influence from the change agent, pharmaceutical company in this case. Cumulative diffusion curves suggest that the pharmaceutical company plays a much weaker role in accelerating the speed of diffusion when a diffusion dynamics is explored with complex agents, defined as heterogenous agents under stage-dependent degrees of external influence. Although our exploration with groups of doctors with different combination of social and professional integration signifies the importance of interpersonal ties, our analysis also reveals that degree of adoption threshold or individual predisposition to knowledge is crucial for adoption decisions. Overall, our approach brings in fresh insights to the burgeoning policy literature exploring complexity, by providing necessary framework for research translation to policy and practice.

Strategic Management, Evolutionary Economics, and Complex Adaptive Systems

This chapter sketches a strategic map of a selection of the relevant issues at the intersection of economics, psychology, sociology, and evolutionary theories applied to strategic management. It takes an evolutionary complexity perspective, based on a (manageable) selection of the relevant literature. The discussion focuses on evolutionary processes of change and their implications for strategic planning and related issues of organisation. The chapter concludes by discussing practical and research issues.

Adaptive Neurodynamics

Neurodynamics is the application of Dynamical system’s theory (DST) to the analysis of the structure and function of recurrent neural networks (RNNs). In this chapter, we present recurrent neural networks artificially evolved for the control of autonomous robots (Evolutionary Robotics), and further analysed within dynamical system’s tenets (Neurodynamics). We search for the characteristic dynamical entities (e.g. attractor landscapes) that arise from being-environment interactions that underpin the adaptation of animat’s (biologically inspired robots). In that way, when an efficient controller is evolved, we are able to pinpoint the reasons for its success in terms of the dynamical characteristics of the evolved networks. The approach is exemplified with the dynamical analysis of an evolved network controller for a small robot that maximizes exploration, while controlling its energy reserves, by resorting to different periodic attractors. Contrasted to other approaches to the study of neural function, neurodynamics’ edge results from causally traceable explanations of behavior, contraposed to just correlations. We conclude with a short discussion about other approaches for artificial brain design, challenges, and future perspectives for Neurodynamics.