Narrow Results

Effective management of supply chains creates value and can strategically position companies. In practice, human beings have been found to be both surprisingly successful and disappointingly inept at managing supply chains. The related fields of cognitive psychology and artificial intelligence have postulated a variety of potential mechanisms to explain this behavior. One of the leading candidates is reinforcement learning. This paper applies agent-based modeling to investigate the comparative behavioral consequences of three simple reinforcement learning algorithms in a multi-stage supply chain. For the first time, our findings show that the specific algorithm that is employed can have dramatic effects on the results obtained. Reinforcement learning is found to be valuable in multi-stage supply chains with several learning agents, as independent agents can learn to coordinate their behavior. However, learning in multi-stage supply chains using these postulated approaches from cognitive psychology and artificial intelligence take extremely long time periods to achieve stability which raises questions about their ability to explain behavior in real supply chains. The fact that it takes thousands of periods for agents to learn in this simple multi-agent setting provides new evidence that real world decision makers are unlikely to be using strict reinforcement learning in practice.

We have studied various spatial models of temporal evolution of opinions of a population of agents, located in a finite closed space, in which, at each time step, a given agent adopts, subject to a probability condition, a new opinion which depends upon the opinions of some of its neighbors. Moreover, at each time step each agent moves to a random neighboring location distributed according to a normal distribution with zero mean and a small standard deviation. The purpose of the paper is to find which types of interactions reduce the number of extremists.

We present a polarity-driven activator-inhibitor model of budding yeast in a two-dimensional medium wherein impeding metabolites secretion (or growth inhibitors) and growth directionality are determined by the local nutrient level. We found that colony size and morphological features varied with nutrient concentration. A branched-type morphology is associated with high impeding metabolite concentration together with a high fraction of distal budding, while opposite conditions (low impeding metabolite concentration, high fraction of proximal budding) promote Eden-type patterns. Increasing the anisotropy factor (or polarity) produced other spatial patterns akin to the electrical breakdown under varying electric field. Rapid changes in the colony morphology, which we conjecture to be equivalent to a transition from an inactive quiescent state to an active budding state, appeared when nutrients were limited.

Long-range interactions are introduced to a two-dimensional model of agents with time-dependent internal variables e_i = 0, ±1 corresponding to valencies of agent emotions. Effects of spontaneous emotion emergence and emotional relaxation processes are taken into account. The valence of agent i depends on valencies of its four nearest neighbors but it is also influenced by long-range interactions corresponding to social relations developed for example by Internet contacts to a randomly chosen community. Two types of such interactions are considered. In the first model the community emotional influence depends only on the sign of its temporary emotion. When the coupling parameter approaches a critical value a phase transition takes place and as result for larger coupling constants the mean group emotion of all agents is nonzero over long time periods. In the second model the community influence is proportional to magnitude of community average emotion. The ordered emotional phase was here observed for a narrow set of system parameters.

Recent analysis of a Yard–Sale (YS) exchange model supplemented with redistributive proportional taxation suggested an asymptotic behavior $P(w)\sim 1∕w^{\mu }$ for the wealth distribution, with a parameter-dependent exponent $\mu$. Revisiting this problem, it is here shown analytically, and confirmed by extensive numerical simulation, that the asymptotic behavior of $P(w)$ is not power-law but rather a Gaussian. When taxation is weak, we furthermore show that a restricted-range power-law behavior appears for wealths around the mean value. The corresponding power-law exponent equals 3/2 when the return distribution has zero mean.

Tuberculosis (TB) is among the 10 top causes of deaths worldwide, and one-quarter of the world population hosts latent TB pathogens. Therefore, avoiding the emergence of drug-resistant strains has become a central issue in TB control. In this work, we propose a nested model for TB transmission and control, wherein both within-host and between-host dynamics are modeled. We use the model to compare the effects of three types of antibiotic treatment protocols and combinations thereof in an in silico population. For a fixed value of antibiotics clearance rate and relative efficacy against resistant strains, the oscillating intermittent protocol, pure or combined, is the most effective against the sensitive strains. However, this protocol also creates a selective advantage for the resistant strains, returning the worst result in comparison to the other protocols. We suggest that nested models should be further developed, since they might be able to inform decision-makers regarding the optimal TB control protocols to be applied under the specific parameters and other epidemiological factors in different populations.

In this work, we study the transmission of the new coronavirus, SARS-CoV-2, which causes COVID-19. Our main aim is to analyze the disease prevalence when vaccination and social distancing strategies are used. Simulations are implemented using an agent-based model (ABM) adapted from a Susceptible-Exposed-Infectious-Recovered (SEIR) type compartmental model. Several scenarios are simulated using the most common vaccines available in Brazil. On each scenario, different fractions of the population are affected by vaccination and social distancing measures. Results show the importance to start public health interventions to reduce the size of the epidemic. Besides, simulations show that vaccination only is not capable to control the disease spread.

Active walk is a paradigm for self-organization and pattern formation in simple and complex systems, originated by Lam in 1992. In an active walk, the walker (an agent) changes the deformable landscape as it walks and is influenced by the changed landscape in choosing its next step. Active walk models have been applied successfully to various biological, chemical and physical systems from the natural sciences, and to economics and many other systems from the social sciences. More recently, it has been used to model human history. In this review, the history, basic concepts, formulation, theories, applications, new developments and open problems of active walk are summarized and discussed. New experimental, theoretical and computer modeling results are included.

Multi-agent systems and agent-oriented methodologies support analysis, characterization and development of complex software systems. These methodologies introduce different definitions for the essential components of multi-agent systems and cover different phases of the system development life cycle. Therefore, appropriate frameworks for evaluation and comparison of different methodologies would support developers to adopt the best methodology, or a combination of different methodologies, based on the project requirements. This review covers the system development phases and the main conceptual components in the context of multi-agent systems. Then, the evaluation frameworks proposed in the literature for comparison and evaluation of agent-oriented methodologies are reviewed. Evaluation frameworks proposed in the literature are categorized into three categories: methodology-based, phase-based and feature-based evaluation frameworks. The paper concludes with the agent-oriented methodologies’ usage challenges, their current limitations and potential future directions.

Group behavior emergent from the systems composed of two types of agents are investigated. The agents are defined on a two-dimensional grid system and move under the influence of the attractive and/or repulsive interactions. Depending on the intensity and the sense of the interactions, a wide variety of spatiotemporal patterns emerge on the system. Those patterns are discussed in terms of the well-known phenomena in real systems such as the residential segregation in cities, cell sorting in multicellular system, self-running droplet, group behavior of a fish school under the attack of a predator and the fission in a cell division process.

Motivated by the desire to bridge the gap between the microscopic description of price formation (agent-based modeling) and the stochastic differential equations approach used classically to describe price evolution at macroscopic time scales, we present a mathematical study of the order book as a multidimensional continuous-time Markov chain and derive several mathematical results in the case of independent Poissonian arrival times. In particular, we show that the cancellation structure is an important factor ensuring the existence of a stationary distribution and the exponential convergence towards it. We also prove, by means of the functional central limit theorem (FCLT), that the rescaled-centered price process converges to a Brownian motion. We illustrate the analysis with numerical simulation and comparison against market data.

In this paper, we present an upper human body tracking system with agent-based architecture. Our agent-based approach departs from process-centric model where the agents are bound to specific processes, and introduces a novel model by which agents are bound to the objects or sub-objects being recognized or tracked. To demonstrate the effectiveness of our system, we use stereo video streams, which are captured by calibrated stereo cameras, as inputs and synthesize human animations which are represented by 3D skeletal motion data. Different from our previous researches, the new system does not require a restricted capture environment with special lighting condition and projected patterns and subjects can wear daily clothes (we do NOT use any markers). With the success from the previous researches, our pre-designed agents are autonomous, self-aware entities that are capable of communicating with other agents to perform tracking within agent coalitions. Each agent with high-level abstracted knowledge seeks 'evidence' for its existence from both low-level features (e.g. motion vector fields, color blobs) as well as from its peers (other agents representing body-parts with which it is compatible). The power of the agent-based approach is the flexibility by which domain information may be encoded within each agent to produce an overall tracking solution.

For business applications of agent-based modeling and simulation, we should formulate rules for individual agents that are more realistic than those required by the KISS principle. This idea could be referred to as "empirical agent-based modeling." In order to develop such models, the first step could possibly be to apply well-established behavioral modeling to individuals/organizations and then link them incorporating interactions between those agents. As an illustrative work, we will present the modeling of TV viewing behavior including interactions in each family, each demographic segment and each reference group.

The conventional wisdom derived from the two-step flow theory suggests that opinion leaders have great influence on their followers. However, it has been difficult for social scientists to measure and describe the extent to which political opinion leaders influence voters, especially when voters today access multiple information sources like communication networks and self-selected news media. This paper fills this gap by using agent-based modeling to represent what the two-step flow theory describes about opinion leader influence and refines the theory based on the findings. First, opinion leader influence does not diffuse to the public without homogeneous communication networks. Second, opinion leader influence usually does not diffuse widely to the public because it inevitably faces resistance from self-strengthening communication networks.

Agent-based modeling is being increasingly used to simulate socio-techno-ecosystems that involve social dynamics. Humans face constraints that they sometimes wish to challenge, and when they do so, they often trigger changes at the scale of the social group too. Including such adaptation dynamics explicitly in our models would allow simulation of the endogenous emergence of rule changes. This paper discusses such an approach in an institutional framework and develops a sequence that allows modeling of endogenous rule changes. Parts of this sequence are implemented in a NetLogo KISS model to provide some illustrative results.

We build an agent-based computational model to study how the changing number of active product variants in a two-dimensional product space affects the performance of different firm types (i.e. large-scale and small-scale enterprises). We use an alternative approach to measure product space dimensionality, considering that dimensions may be a fraction of the Euclidean measure. The results confirm that high dimensionality gives advantage to small-scale firms. Additionally, we find that large-scale firms may also benefit from initial increasing dimensionality, since it allows a small degree of product differentiation and price discrimination.

A lot of agent-based models were built to study diffusion of innovations. In most of these models, beliefs of individuals about the innovation were not represented at all, or in a highly simplified way. In this paper, we argue that representing beliefs could help to tackle problematics identified for diffusion of innovations, like misunderstanding of information, which can lead to diffusion failure, or diffusion of linked inventions. We propose a formalization of beliefs and messages as associative networks. This representation allows one to study the social representations of innovations and to validate diffusion models against real data. It could also make models usable to analyze diffusion prior to the product launch. Our approach is illustrated by a simulation of iPod™ diffusion.

Exploiting a precise reproduction of a stock exchange, the robustness of the continuous double auction (CDA) mechanism, evaluated by means of the waiting time distributions, has been proved versus 36 different setups made by varying both the operators' behavior and the market micro structure. The obtained results demonstrate that the CDA remains able to clear strongly different order flows, although the Milan stock exchange seemed to be a little more efficient than the NYSE under the allocative point of view, evidencing the intrinsic complexity of the stock market. The simulation has been built as an agent-based model in order to obtain a plausible order flow. The decisions of single agents and their interaction through the market book are realistic and reproduce some empirical analysis results. The mentioned results have been obtained either by the analysis of the complete pending time series or the same computation of the asks and bids series alone.

We present a model for evacuees' exit selection in emergency evacuations. The model is based on the game theoretic concept of best-response dynamics, where each player updates his strategy periodically by reacting optimally to other players' strategies. A fixed point of the system of all players' best-response functions defines a Nash equilibrium (NE) of the game. In the model, the players are the evacuees and the strategies are the possible target exits. We present a mathematical formulation for the model and show that the game has a NE with pure strategies. We also analyze different iterative methods for finding the NE and derive an upper bound for the number of iterations needed to find the equilibrium. Numerical simulations are used to analyze the properties of the model.

We report on replications of early experiments with FEARLUS, using larger numbers of agents, larger numbers of land parcels, and greater network connectivity than in the original work. We find that results from the larger-scale experiments differ from the smaller environments used previously. Whilst results from small communities of agents and environments should not be ignored just because the same effect is not observed in larger communities (and indeed vice versa), this work does raise the extent to which more general conclusions can be drawn from agent-based studies involving fixed population or environment sizes.