Ecology Of Agents Research Articles

The Many Faces of Agent-Based Computational Economics: Ecology of Agents, Bottom-Up Approaches and Technical Breakthrough.

This paper presents an overview of how agent-based computational economics (ACE) can contribute to the study of economic systems both at the macro and the micro level. It highlights the way these models can improve our understanding of social interactions and coordination mechanisms and bring into light the complex dependencies between the micro and the macro level. In a first part, the differences and the complementary nature between ACE and other quantitative methods (orthodox and coming from other disciplines) are underlined. A second part presents the importance of ACE for a better understanding of the functioning of markets. ABM allows to simulate heterogeneous agents and different types of interactions, between individuals, between individuals and institutions, between institutions. The specific characteristics of the goods are identified and the market design can vary. Based on the analysis of various markets, this discussion brings fresh insights to a broader and very long-standing debate about the conditions of efficiency of market structures.

Experiential Learning, Competitive Selection, and Downside Risk: A New Perspective on Managerial Risk Taking

It is part of managerial wisdom that managers need to take risks in order to succeed. This is in stark contrast to the prominent agent-based theories of experiential learning and competitive selection that are known to be biased against risky alternatives in the long run. How can the positive managerial view on risk taking prevail given these results? Qualitative surveys of managers suggest risks to be acceptable if their outcome distribution meets certain criteria. We argue in this paper that these criteria are widely congruent with low downside risk. Using a novel simulation design, we analyze the effects of experiential learning and competitive selection on downside risk preferences in an ecology of agents. In the long run, experiential learning implies weak downside risk seeking, whereas competitive selection leads to strong downside risk aversion. Furthermore, competitive selection leads to the prevalence of outcome distributions with low downside risk in an ecology, even if they have a significantly higher level of total risk than comparable distributions with more downside risk. We draw implications for empirical studies on managerial behavior. The online appendix is available at https://doi.org/10.1287/orsc.2017.1149 .

Asset Pricing with Heterogeneous Investment Horizons

We consider an analytically tractable asset pricing model describing the trading activity in a stylized market with two assets. Traders are boundedly rational expected utility maximizers with different beliefs about future prices and different investment horizons. In particular, we analyze the effects of the latter source of heterogeneity on the dynamics of price. We find that in the case with homogeneous agents, longer investment horizons lead to more stable dynamics. This is not true, however, in the case of a mixed population of traders, when the increase of heterogeneity in the investment horizons can introduce instability in the system. Furthermore, the role of heterogeneity turns out to be different for different trading behaviors and its effect on the aggregate dynamics depends on the whole ecology of agents' beliefs.

Molecular characterization and evaluation of virulence factors of Cryptococcus laurentii and Cryptococcus neoformans strains isolated from external hospital areas

Cryptococcosis is a common opportunistic fungal infection that is mainly caused by the species Cryptococcus neoformans and Cryptococcus gattii, but there have recently been several reports of infection by non- neoformans Cryptococcus species. The aims of this study were to genetically characterize Cryptococcus spp. isolated from external hospital areas in Minas Gerais State, Brazil, and to evaluate their pathogenic potential, analyzing their phospholipase and melanin production and the capacity for capsule enlargement. Seventy-three different samples were collected: 62 from bird droppings and 11 from tree detritus. C. neoformans alone was isolated from 43.8 % of the samples, Cryptococcus laurentii alone from 23.3 % and both fungi were found together in 10.9 %. C. laurentii was exclusively isolated from 45 % (5/11) of the tree samples ( Anacardium occidentale, Guazuma ulmifolia, Mangifera indica and Ficus benjamina). Among the 51 C. neoformans isolates, 47 were classified as type VNI and four as type VNII. All of the C. neoformans isolates were of MATα type. Among the 21 isolates of C. laurentii genotyped using the URA5-RFLP technique, 16 amplified a 1.6 kb amplicon which produced a specific restriction profile in 15 isolates. In C. neoformans, 76.4 % of the isolates were capable of capsule enlargement in the induction medium and 92.1 % were phospholipase producers. In C. laurentii, 7.4 % of the isolates were capable of capsule enlargement and 85.1 % were phospholipase producers. Characterization of the genotypes and the pathogenic potential of the Cryptococcus spp. isolates studied may contribute towards better understanding of the epidemiology of cryptococcosis and the ecology of agents causing this disease in our region.