Agent-Based Approaches in Economic and Social Complex Systems VIII

In the context of political modernization and economic development, the complex adaptive systems framework can help address the coupling of macro social constraint and opportunity with individual agency. Using a simple evolutionary game approach, we fuse endogenously derived socio-economic system dynamics from Human Development (HD) Theory with Prisoner’s Dilemma spatial intra-societal economic transactions. We then explore a new human development dynamics (HDD) model behavior via quasi-global simulation methods to explore technological progression on economic development, cultural plasticity, social and political change. Using network analysis, we then investigate the impact of technology proliferation on communications ease and the resulting compression of social space on individual wealth and political preference formation. As economic and social capital is created, past transaction histories tend to reinforce future success, and networks emerge and solidify at different rates depending on technology. Increasing social connectivity in small populations has an immediate and positive impact on wealth creation, yet those effects become negative as technology proliferates and population size increases. This suggests not only diminishing marginal returns to increasing communications’ payoffs to individuals but moreover crowding out effects. We believe complex adaptive or evolutionary systems approaches are necessary to understand both near and potentially catastrophic, far-from-equilibrium behavior and societal outcomes across all human scales of modernization.

Research into the cognitive origins of prejudice has largely focused on individual psychological processes. We introduce a novel Agent-based model that simulates both individual-level and inter-personal processes, and that allows exploring how stereotypes are shared and validated through interpersonal processes. At the individual-level, agent processes are simulated by recurrent auto-associative networks. To simulate interpersonal processes, these individual networks are combined into a “community of networks” so that they can exchange their individual information with each other by transmitting information on the same concepts from one network to another. Through simulations, it is shown how the model can account for a number of seminal findings from the empirical literature on illusory correlations, a key cognitive antecedent of prejudice. In addition, novel hypotheses in terms of the impact of interpersonal processes on the dissemination of illusory correlations were supported by the results of a small group experiment. We discuss the results and argue that agent-based models can provide a first step in integrating individual and interpersonal processes underlying stereotype formation and IC.

We constructed an agent-based model (ABM) that tested how globalization, frequent movements of individuals between local societies, affects the accumulation of cultures. In the model, multiple groups were connected as a circular stepping-stone formation without boundaries. Agents copy cultural traits of others in their groups; agents may gain or lose their cultural traits through the process, depending on the traits of opponents, which is called within-boundary communication. Agents periodically migrate between adjacent groups. Agents also visit adjacent groups, copy the cultural traits in those groups, and return to their group, which is called cross-boundary communication. The model indicates that cultural traits may accumulate in the whole population even if they migrate frequently. The necessary conditions are that agents also frequently communicate cross-boundary by finding an appropriate group.

Category theory and, more specifically, topos categories provide a more expressive type of mathematical modeling and, thereby, open the door to social models that are both rigorous and expressive. The present analysis draws upon four views of topos categories to construct a rich model of a logic based on recognition theory. The topos initially provides support for: (1) set theory (and classical logic), (2) topological regions dualism. The resulting structures support: (3) a local intuitionist logic that can vary by situated circumstances and actor types, and (4) the specification of regions that are defined by finely-differentiated classifiers. These tools construct an integrated topos category of social recognition that supports diverse forms of local logic. The latter are then explored regarding their contributions to a mathematical model of social conflict. Categorial analysis of historical patterns, as well as a complementary simulation model, is used to illustrate the advantages of such an approach.

The various reputation theories that proposed to solve the problem of social order have an unrealistic assumption that all agents observe what occurred in all other agents. To improve this situation, Nakai and Muto (2008) proposed us-Tit For Tat (TFT) strategy that requires an agent to regard as a friend one who did not attack himself/herself and his/her “friends”, and they found a resultant peaceful state. They assume that the us-TFT agent must observe only what occurred in himself/herself and his/her friends. However, when all agents become mutual friends, they observe all other agents, hence us-TFT suffers from the same weakness of related studies. To overcome the weakness, we propose a new us-TFT strategy with which an agent observes a limited number of other agents. A limited number of other agents are selected based on how peacefully they act toward the new us-TFT agent. We performed evolutionary simulations with this strategy and found the emergence of a peaceful state. It means that a peaceful state can emerge without observing all other agent’s actions. In addition, we find that the more the limited number is, the less frequently a peaceful state emerge.

Different from social network of practice that emphasizes relations among members with weak ties, community-based learning more focuses on the competence and practice of individuals connected by strong ties. With respect to the E-government service adoption, we assume that learning within communities is more common than that via social network. On the other side, the spread of information/knowledge on E-government might also influence the learning process, and further affect the adoption behaviour indirectly. Understanding such dynamic learning mechanism is crucial to the investigation of divergent citizens’ adoption behaviour of E-government services, thus potentially important to the evaluation and design of supporting policies as well. In order to investigate the influence of learning within communities that are composed of citizens with different characteristics, and to explore the effectiveness of supporting policies in a long-term perspective, agent-based modelling is applied. This model could enable the understanding of a wide range of possible adoption behaviours under different scenarios, and the exploration of to what extent the variant supporting policies are effective.

This paper proposes a method for analyzing architectural space design based on human behavior semiosis by using a multiagent simulator as a design aid. Face-to-face interviews with those who visited a campus plaza were conducted as a case study in order to understand the processes behind human behavior, and the space design based on the interview analysis is considered. This study is organized as follows: (1) a human behavior semiosis model is proposed based on C.S. Peirce’s concept of semiosis; (2) an analysis of the interview results based on the model of step 1 is proposed; (3) the site is modeled using cells in a multiagent simulator; and (4) based on the analysis of step 2, the modeled site is evaluated according to the human behavior semiosis performed by the multiagent simulator in step 3.

Vaccination subsidy is one of the major types of public health services. In order to evaluate the efficiency of the subsidy, this research carries out an Agent-Based Simulation approach. By leading strategies with different levels of subsidy into the social system and simulating the spread of influenza infection on a layout based on the daily life of every agent, the research constructs an immunization model depending on decision making of each agent. As a result, the model suggests that vaccination subsidy tend to take a positive role and gives decision support to assess the relative impact of public health services for influenza control.

An agent-based model of the investment game is proposed to study the complex dynamics between trust, growth, and inequality with different underlying technologies. It is found that agents in this economy, through learning to trust and to be trustworthy, are able to coordinate themselves well in networking, which hence facilitates wealth creation. The excessive smoothness in economic growth is, therefore, prevalent in all simulations, and the underlying technologies can only determine the speed of growth and network formation. While the advancement of technology can ameliorate the inequality of wealth distribution, it also lowers the social mobility of the agents.

Players are motivated by social norms and morale to exert extra efforts in a work place with high morale. However, we often observe that the initially high performance of an organization gradually declines. In order to seek for an optimal wage incentive system, we have constructed an agent-based model with a small group of heterogeneous or homogeneous workers. The model has the following features: players’ behavior is subject to random shocks; inertia effect is introduced; and players are rewarded on the basis of their performances. The virtual experiment demonstrates that an organization, consisting of homogenous players, is more viable against the erosion than that with heterogeneous players, and that an organization with homogenous players is far more vulnerable to the erosion than that with heterogeneous players when subject to random shocks. Interestingly, an organization with heterogeneous agents can enjoy high morale among players that allows them to maintain high performance. We also show that the reward incentive is remarkably effective in any organization and it is particularly powerful in homogenous organizations. We have compared the two personnel rating systems: one based upon the level of performance relative to other workers’ contemporaneous ones and the other upon her own past performance. In inducing higher performance of the workers, the latter system excels the former for the heterogeneous organization, whereas the former system will outperform for the homogeneous one.

In this Study, we design a multi agent system within various agents such as enterprise agents, investor agents, and customer agents in our model. Enterprise agents using stock as collateral for loans from banks, they use these resources to execute business activities such as purchase and production, and sell products to consumers to gain profit, part of profit will become dividends to investors.

We design a simple model of enterprises’ financial statements. Not only technical analysis but also fundamental analysis will be implemented as measures of investments in our model. We will clear up influence of trading strategies of investor agents in the artificial market to business managements in our agent based model.

As the results, measures of investments of investors in a stock market influence on business activities in our rational assumptions. Totally, the environments exert healthy influence on business activities are trend strategy dominating markets and random strategy dominating markets. The key point is a stable stock price or a good predictability of stock price.

In this chapter, we examine the empirical relationship between return volatility and behavioral heterogeneity in a structural stochastic volatility model. First, we study the empirical performance of the model with two trading mechanisms via moment-based estimation from S&P 500. Second, we compare the empirical performance of both specifications. In particular, a simulated test distribution is used to evaluate the significant difference between the two models. The result of the formal test shows that the model incorporating herding fits the data better than the model incorporating wealth, but that they are not significantly different at the 5 % level.

We present a method to develop agent-based models based on the service system abstraction and the ISPAR model, to study market systems in the light of Service-Dominant logic. Service-Dominant logic is a mindset that acts as a lens, through which we can see the world in a different way. More precisely, it intends to bring new insights into the marketing domain by perceiving the market as a system of resource integrating actors, who exchange resources in the form of services. The service system abstraction and the ISPAR model of service system interactions provide a reasonable starting point for this endeavor. We explain how the service system abstraction and the ISPAR model could be implemented computationally through an agent-based model of a hypothetical market system. We further argue that the proposed method could be successfully applied to study various real market phenomena.

We showed the fact that market mavens are included more among communicative consumers than average and act as information hubs between communities by analyzing a research about consumers’ word-of-mouth activities using snowball sampling technique. We also designed research questions to seize the feature of community structure inside social networks. On the basis of these outcomes, we developed a network model that reflects community structure and a multi-agent simulation model. The agreement between research and simulation results endorsed the factors premised in the model such as four consumer types, type dispositions on network, and effect of local diffusion. Moreover, we used this simulation model to propose a method for improving the effect of a referral program. An advantage of this method is that marketers need not to have special information about individual customers.