Добірка наукової літератури з теми "Imperfect Selection"

In this paper we analyze players' long-run behavior in evolutionary coordination games with imperfect monitoring in a large population. Players can observe signals corresponding to other players' unseen actions and use the proposed simple or maximum likelihood estimation algorithm to extract information from the signals. In the simple learning process we find conditions for the risk-dominant and the non-risk-dominant equilibria to emerge alone in the long run. Furthermore, we find that the two equilibria can coexist in the long run. In contrast, the coexistence of the two equilibria is the only limit distribution under the maximum likelihood estimation learning algorithm. We also analyze the long-run equilibria of other 2x2 symmetric games under imperfect monitoring.

Abstract This paper studies how the congestion of its editorial process affects an academic journal. In a publishing game played by researchers and editors, we assume that quality screening by editors depends on their ability to properly process the flow of submissions. When too many papers arrive, this ability declines and editors may reject good papers or accept papers with little contribution to scientific knowledge. In this game, a separating equilibrium always exists in which only good researchers submit their papers to the journal. Each paper is accepted and the quality of the journal reaches its highest level. However, when the researchers’ reward for each publication exceeds a given threshold, two hybrid equilibria are also feasible. In these equilibria, authors of low-quality papers submit their papers to the journal and, as the flow of papers exceeds the editors’ capacity of perfect assessment, the selection process becomes imperfect. This creates an opportunity for authors to submit poor quality papers, a behaviour which contributes to the congestion of the editorial process. The various strategies implemented by the editors to oppose congestion are then discussed.

In this paper we analyze players' long-run behavior in evolutionary coordination games with imperfect monitoring in a large population. Players can observe signals corresponding to other players' unseen actions and use the proposed simple or maximum likelihood estimation algorithm to extract information from the signals. In the simple learning process we find conditions for the risk-dominant and the non-risk-dominant equilibria to emerge alone in the long run. Furthermore, we find that the two equilibria can coexist in the long run. In contrast, the coexistence of the two equilibria is the only limit distribution under the maximum likelihood estimation learning algorithm. We also analyze the long-run equilibria of other 2x2 symmetric games under imperfect monitoring.

Abstract Mimicry is widely used to exemplify natural selection’s power in promoting adaptation. Nonetheless, it has become increasingly clear that mimicry is frequently imprecise. Indeed, the phenotypic match is often poor between mimics and models in many Batesian mimicry complexes and among co-mimics in many Müllerian mimicry complexes. Here, we consider whether such imperfect mimicry represents an evolutionary compromise between predator-mediated selection favoring mimetic convergence on the one hand and competitively mediated selection favoring divergence on the other hand. Specifically, for mimicry to be effective, mimics and their models/co-mimics should occur together. Yet, co-occurring species that are phenotypically similar often compete for resources, successful reproduction, or both. As an adaptive response to minimize such costly interactions, interacting species may diverge phenotypically through an evolutionary process known as character displacement. Such divergence between mimics and their models/co-mimics may thereby result in imperfect mimicry. We review the various ways in which character displacement could promote imprecise mimicry, describe the conditions under which this process may be especially likely to produce imperfect mimicry, examine a possible case study, and discuss avenues for future research. Generally, character displacement may play an underappreciated role in fostering inexact mimicry.

Spanish
Ph.D.
One of the most researched challenges in learning Spanish is the acquisition and use of the past aspect, namely, the preterit and imperfect. L2 learners encounter this challenge due to differences in how native English and Spanish speakers view past events. Numerous studies on the Spanish past aspect have analyzed L2 learners’ past aspectual selections through two hypotheses: the lexical aspect hypothesis (LAH), which claims that lower-level L2 learners are guided by the lexical semantics of the verb in their selections of past aspect; and the discourse hypothesis (DH), which claims that as L2 learners become more proficient, they make past aspectual selections to foreground and background information. The present study uses both hypotheses to analyze past aspectual selections in beginner, intermediate, and advanced L2 learners as well as native speakers. By doing so, it was possible to analyze how past aspectual selections differ across proficiency levels on the basis of lexical aspect and narrative grounding. 75 L2 learners and 20 native Spanish speakers produced two uncontrolled, written narratives. The first was a film-retell based on a five minute clip of “Alone and Hungry” from Modern Times and the second was a personal narration of a favorite vacation. All verbs within each narrative were coded for lexical aspect: atelic states and activities as well as telic accomplishments and achievements. Next, foregrounded clauses (preterit) were separated from backgrounded clauses (imperfect) in order to capture how narrative structure played a role in past aspectual selections. The researcher and a second coder determined the accuracy of use of the preterit and imperfect. Quantitative data consisted of contingency tables and chi-square analyses for the film-retell task and the personal narrative task separately that captured total use of preterit and imperfect morphology for the LAH. It also captured correct use of preterit and imperfect for the DH. Finally, type-token ratios (TTRs) were used to assess the lexical variety of verbs on the basis of grammatical aspect, lexical aspect, and proficiency level for the fill-retell task. The results revealed that for the film-retell and personal narrative tasks, there was support for the LAH. Participants across proficiency groups were guided by the LAH when making past aspectual selections. There was also partial support for the DH in that participants across proficiency groups consistently selected preterit morphology for atelic states and activities, even if the lower-level proficiency groups made more errors in their past aspectual selections. However, showed very little evidence of correct use of imperfect morphology with telic achievements and accomplishments. Additionally, analysis of type-token ratios (TTRs) showed that participants across proficiency groups repeatedly used high-frequency verbs to complete their narratives. This was particularly noticeable for state and activity verbs within the beginner group. Task type appeared to be a major influence in participants’ selections of past aspect for the film-retell task. Participants were influenced by the numerous sequential scenes of the film task, which accounted for the high frequencies of preterit to mark bounded events, but much lower frequencies of imperfect to mark unbounded events. For the personal narrative, frequencies of the preterit were higher than imperfect across each lexical class. When assessing both film-retell and personal narratives, participants produced very few contexts for imperfect with achievement and accomplishment verbs. Participants either did not have the knowledge to do so or felt that their own narrations of “Alone and Hungry” and their personal narrative was sufficient to complete the task.
Temple University--Theses

Thesis (Ph. D.)--Ohio State University, 2005.
Title from first page of PDF file. Document formatted into pages; contains xi, 124 p.; also includes graphics. Includes bibliographical references (p. 121-124). Available online via OhioLINK's ETD Center

This dissertation is a collection of essays (chapters) on behavioral economics. Behavioral economics—arguably one of the most influential innovations in economics over the last 20 years—is a research paradigm introducing psychologically more realistic assumptions into economics. A common theme throughout the dissertation is the focus on either biased beliefs, or heterogeneous preferences, or both. The first chapter serves as an introduction to some themes in behavioral economics and its implications for market outcomes in industrial organization settings. The next two chapters are theoretical papers on entrepreneurial and managerial overconfidence that can also be thought of as contributions to this newly emerging field of behavioral industrial organization. The last chapter is an empirical contribution on gender differences in preferences and economic behavior at a young age.

Diss. Stockholm : Handelshögskolan, 2011

Cette thèse présente de nouveaux résultats théoriques et empiriques sur les propriétés de la dynamique des entreprises, de l'industrie et du commerce international, et vise à répondre à une série de questions de recherche liées à l'explication de l'hétérogénéité des entreprises dans le contexte de l'apprentissage et de la sélection imparfaite des entreprises.Premièrement, nous proposons une enquête de la littérature sur les mécanismes de sélection de marchés. Nous regroupons les travaux de recherche sur ce thème en trois paradigmes théoriques, que nous concilions en termes de lois de sélection. Nous montrons que les trois paradigmes ont convergé vers des mécanismes de sélection fondé sur l'hétérogénéité des entreprises et sur les rendements croissants, qui reposent cependant sur des théories qui diffèrent en termes de sources de rendements croissants, de mécanismes de génération d'hétérogénéité des entreprises, de rationalité des entreprises, et pour l'accent mis sur les états d'équilibre vis-à-vis de la dynamique hors équilibre.Deuxièmement, nous construisons un modèle de commerce international à deux pays avec des entreprises hétérogènes pour étudier les effets que l'apprentissage technologique des entreprises et la sélection imparfaite des marchés exercent sur les flux d'exportation, les parts de marché et les productivités des entreprises. Dans ce modèle nous faisons l'hypothèse que la selection de marché dans chaque pays est guidée par un processus d'urnes de Pólya fini, qui incorpore des rendements croissants dynamiques au niveau de l'entreprise. Nous montrons que, en présence d'une distribution statique de la productivité des entreprises, le processus de sélection du marché conduit au monopole. Lorsque que l'apprentissage de l'entreprise est inclus dans le modèle, les marchés convergent vers des structures non monopolistiques, dont le degré de concurrence dépend de l'ouverture commerciale et de l'intensité de la sélection. Enfin, nous montrons que notre modèle est capable de reproduire conjointement un large ensemble de faits stylisés concernant les flux d'exportation intra-industriels, l'industrie et la dynamique des entreprises. De plus, nous montrons que une augmentation de l'intensité de la competition sur les marchés internationaux augmente la concentration et la volatilité.Troisièmement, nous étudions la relation entre la volatilité des taux de croissance et la taille des entreprises et ses déterminants en utilisant une base de données sur les entreprises manufacturières françaises entre 1993 et 2009. Nous montrons que cette rélation, au niveau agrégé et sectoriel, s'écarte de l'approximation linéaire trouvée dans les études empiriques précédentes. Elle est en effet en forme de J, très raide pour les petites entreprises et plate pour les grandes. Ensuite, nous expliquons ce nouveau résultat empirique à l'aide d'un un modèle de sélection imparfaite axé sur la concurrence des entreprises sur la base de la taille et de la productivité.Enfin, nous étudions les caractéristiques des entreprises qui utilisent l'intelligence artificielle (IA) et le lien entre l'utilisation de l'IA et la productivité à l'aide d'une étude empirique reposant sur une base de données des entreprises françaises en 2018. Nos resultats montrent que les utilisateurs de l'IA ont tendance à être plus grands et plus jeunes que les non-utilisateurs. De plus, l'utilisation de l'IA est positivement liée à plusieurs facteurs complémentaires, comme le capital humain ou la présence d'infrastructures et technologies numériques. Ensuite, nous montrons que les utilisateurs d'IA avec la plus grande taille sont plus productifs, mais que cet advantage est lié à la selection dans d'utilisation de l'IA. Cependant, nous trouvons un lien positif entre l'IA et la productivité pour les développeurs d'IA qui s'étend au-delà de la sélection, en particulier lorsque la croissance de la productivité sur une période suffisamment longue est prise en compte
This dissertation presents new theoretical and empirical evidence on the properties of firms' and industry dynamics and international trade. In particular, this thesis aims at answering a series research questions linked to the explanation of firms' heterogeneity in the context of the most recent findings related to both firms' learning and imperfect selection.First, we provide a survey of the main mechanisms of market selection used in economics. We gather them in three theoretical paradigms, that we try to reconcile in terms of underlying laws of selection. We show that the three paradigms have been converging to selection mechanisms focussing on firm heterogeneity and increasing returns, that are however fostered by theories which differ in terms of sources of increasing returns, generating mechanisms of firm heterogeneity, firm rationality and emphasis on equilibrium states vis-á-vis out-of-equilibrium dynamics. Our discussion suggests that the convergence between the theoretical paradigms is taking place in the direction of research, which is aimed at the replication of empirical patterns related to firm heterogeneity, rather than in the theory underlying selection mechanisms.Second, we build a simple international trade two-country model of competition among heterogeneous firms to study the effects that firm learning and imperfect market selection exert on export flows, market shares and firm productivities. Market selection in each country is driven by a finite pairwise Pólya urn process, which embodies dynamic increasing returns at the firm level. In presence of a static distribution of firm productivity, the market selection process leads to a monopoly. When firm learning is included in the model, markets converge to non-monopolistic structures, whose degree of competition depends on trade openness and selection intensity. Finally, we show that our simple stochastic model with firm learning and imperfect selection is able to jointly reproduce a wide ensemble of stylized facts concerning intra-industry trade, industry and firm dynamics. In addition, we show that trade activities increase concentration and volatility.Third, we investigate the firm growth rates volatility-size relation and its determinants in a comprehensive dataset of French manufacturing firms between 1993 and 2009. Differently from previous contributions, we study the relation using sales data for firms at both the aggregate and sectoral level. First, we show that the relation deviates from the linear approximation found in previous studies. It is indeed J-shaped, very steep for small firms and flat for large ones. Second, we explain this new empirical finding via a tractable model of imperfect selection encompassing firms competing on the basis of both size and productivity. Our contribution suggests that large firms are Gibrat's and that the empirical shape of the firms' growth rates variance-size relation can be explained by imperfect selection mechanisms whose outcomes are mediated by both the strength of shares reallocation and firms' joint heterogeneity in size and productivity.Finally, we investigate the characteristics of firms using Artificial Intelligence (AI) and the link between AI use and productivity in a comprehensive database of French firms in 2018. We find that AI users tend to be larger and younger than non-users. AI use is positively related to several complementary assets, including digital infrastructure, complementary digital technologies, and human capital. Focussing on the AI-productivity nexus, we show that the largest AI users are more productive, but that this premium is related to their selection into AI use. When we consider either all AI users or AI buyers, no average relationship between AI and productivity growth could be retrieved. However, we find a positive AI-productivity link for AI developers, especially when productivity growth over a sufficiently long time period is considered

碩士
國立成功大學
工業與資訊管理學系碩博士班
94
For manufacturing systems, the production technology usually has an influence on the production lot size and therefore delimits the capacity. The technology cost is closely related with the production lot size. This research extends the inventory model, proposed by Khouja (2005), to formulate the models by treating the lot size as the decision variable at the technology selection stage and considering products with imperfect quality. It is assumed that poor-quality items will be sold as a single batch after the 100% screening process. The objective of this research is to determine the optimal production lot size to maximize the total profit. Three inventory models are presented for considering the fixed defective rate, the fuzzy defective rate, and the stochastic defective rate, respectively. For new products, the determination of the defective rate is usually based on the judgment of experts, practically using fuzzy numbers to express experts’ linguistic descriptions. We can use statistical techniques to estimate the defective rate, and use probability distribution to describe the uncertainty of the defective rate, once production data are sufficient. Considering the above scenarios, we not only formulate the deterministic model, but investigate two kinds of uncertain models. The defective rate of a lot size is affected by the quality of raw material, the system’s configurations, the product design, etc. We can use the fuzzy theory to describe the uncertainty of the defective rate when historical production data are insufficient. In the fuzzy model, we employ the Yager’s ranking method to estimate the total profit per unit time, and then derive the optimal production lot size by the total profit function. The decision-makers can use the solutions to select appropriate production technology and production lot size.

碩士
國立交通大學
電信工程研究所
100
The Orthogonal Frequency Division Multiple Access (OFDMA) scheme is an efficient anti-fading transmission scheme which renders high spectral efficiency and simple channel equalization. It also allows °exible resource allocation (RA) to meet various user requirements and achieve maximum network capacity. With the help of relays, link quality at cell edge can be improved and both network capacity and the coverage area can therefore be improved. In this thesis, we consider the problem of RA and relay selection for downlink trans- mission in both single-input single-output (SISO) and multiple-input multiple-output (MIMO) OFDMA based cellular networks. We assume the availability of multiple co- operative relay stations but not the perfect channel state information (CSI). Instead, the base station knows only the estimated channel (link) gain and the associated error distribution. We use a tight capacity lower bound (CLB) for a link with imperfect CSI as the performance metric. In SISO networks, we derive the optimal source and relay power allocation ratio that maximizes the CLB of a cascaded source-relay-destination link. Based on this optimal power ratio, we propose a simple suboptimal algorithm that assigns power, subcarriers and cooperative relays to each serving mobile station. We then derive the optimal power ratio for MIMO networks. Using the proposed subcar- rier assignment algorithm for SISO network, we present the optimal and a suboptimal power allocation schemes. To reduce the computation complexity, we derive a near- optimal power ratio, assuming both source-to-relay and relay-to-destination links have the same rank. Simulation results show that our algorithm not only meets the users' rate constraints with very high probabilities but yields an excellent sum rate (CLB) performance.