A connectionist model of the organizational learning curve

Organizational learning can be understood as a spontaneous development of routines. Mathematically, this process can be described as a search for better paths on a graph whose nodes are humans and machines. Since the rules for connecting nodes depend on their ability to process goods, the slope of the learning curve may be connected to physical and psychological properties. Two suggestive examples are discussed. Guido Fioretti, born 1964, graduated in Electronic Engineering and obtained a PhD in Economics from the University of Rome “La Sapienza”. He is currently an assistant professor at the University of Bologna, Italy.His research interests span from decision theory to economics and organization science. In particular, he is interested in linking structural development to cognitive processes. The present article has been conceived as a theoretical underpinning of agent-based simulations of organizations. In particular, future applications of the Java Enterprise Simulator () may test the usefulness of the results derived herein.     

A binary particle swarm optimization algorithm inspired by multi-level organizational learning behavior

Recently, nature-inspired algorithms have increasingly attracted the attention of researchers. Due to the fact that in BPSO the position vectors consisting of ‘0’ and ‘1’ can be seen as a decision behavior (support or oppose), in this paper, we propose a BPSO with hierarchical structure (BPSO_HS for short), on the basis of multi-level organizational learning behavior. At each iteration of BPSO_HS, particles are divided into two classes, named ‘leaders’ and ‘followers’, and different evolutionary strategies are used in each class. In addition, the mutation strategy is adopted to overcome the premature convergence and slow convergent speed during the later stages of optimization. The algorithm was tested on two discrete optimization problems (Traveling Salesman and Bin Packing) as well as seven real-parameter functions. The experimental results showed that the performance of BPSO_HS was significantly better than several existing algorithms.     

A semiparametric additive rates model for clustered recurrent event data

Recurrent event data often arises in biomedical studies, and individuals within a cluster might not be independent. We propose a semiparametric additive rates model for clustered recurrent event data, wherein the covariates are assumed to add to the unspecified baseline rate. For the inference on the model parameters, estimating equation approaches are developed, and both large and finite sample properties of the proposed estimators are established.     

A model for real poker with an upper bound of assets

This paper considers a continuous model of two-person poker, where the maximal amount of betB is assumed and the player who acts first chooses the amount of bet in the game. We analyze a model, in which the range of the amount of bet is a finite interval [0,B], 0B<+, to obtain a saddle point of the payoff function as a pair of optimal strategies among mixed strategies. We compare our results with those of Karlin and Restrepo and those of Newman.The author wishes to thank Professors M. Sakaguchi and T. Kurisu of Osaka University for suggesting the problem as well as for guidance and encouragement.     

The estimation of a nonlinear moving average model

We consider discrete-parameter stochastic processes that are the output of a nonlinear filter driven by white noise. For a simple model, we derive estimates of the unknown coefficients in the transfer function and the noise variance, and investigate their asymptotic properties. We prove some lemmas that can also be used to obtain rates of convergence in the weak and strong laws of large numbers, and central limit theorems, for estimates of more general nonlinear models.     

Holdings of financial assets: A Markov chain analysis

This paper introduces the Markov chain model as a simple tool for analyzing the pattern of financial asset holdings over time. The model is based on transition probabilities which give the probability of switching $1 of wealth from one asset to another. An illustrative application is provided.     

A Model of Obsolescence of Organizational Rules

The phenomenon of obsolescence has found little attention in past research on organizational change. Most research on organizational change has instead been concerned with the opposite of obsolescence, that is, with mechanisms which render organizational structures increasingly persistent. However, everyday experience teaches that organizational structures become outdated as surrounding conditions shift. This article explores obsolescence by means of mathematical modeling and empirical analysis. Obsolescence is conceptualized as a result of flows of conditions in, out, and between two states: consistent conditions and inconsistent conditions. The model predicts that the obsoleteness curve rises with decreasing increments over time. This is tested with empirical data on organizational rule change. The results confirm the obsolescence model. This revised version was published online in August 2006 with corrections to the Cover Date.     

Systemic view of learning scientific concepts: A description in terms of directed graph model

This study suggests a systemic view on concept learning, in which concepts are seen as complex, dynamically evolving structures with robust and persistent patterns emerging from loosely connected elemental pieces. The model is represented in the form of a directed graph, and it explains many empirically identified features of learning paths, including: (1) the stability of learned concepts, (2) changes in concept attributions (ontological shift) as guided by theoretical knowledge, and (3) the evolution of concepts as driven by model competition with the accumulation of evidence. The results thus show applicability of the systemic view and the directed graph model. © 2013 Wiley Periodicals, Inc. Complexity 19: 27–37, 2014     

A dynamic model of the distribution of wealth among households and nations

Stiglitz once showed that, in general, aggregate wealth is asymptotically uniformally distributed among individuals. However, in his formulation household saving is not the outcome of utility maximization over time. Constructing a simple dynamic general equilibrium model in which household saving is choice-theoretically determined, we show that given initial holdings of wealth there is a unique and stable steady state distribution of wealth and that the distribution of wealth becomes more even (resp. uneven) as time goes by if the total wealth is initially greater (resp. smaller) than its steady state level. We also study the response of the steady state equilibrium to the changes in initial distribution of wealth and the rate of time preference, and to several types of technological improvements.     

A semiparametric Bayesian generalized linear mixed model for the reliability of Kevlar fibers

We analyze the reliability of NASA composite pressure vessels by using a new Bayesian semiparametric model. The data set consists of lifetimes of pressure vessels, wrapped with a Kevlar fiber, grouped by spool, subject to different stress levels; 10% of the data are right censored. The model that we consider is a regression on the log‐scale for the lifetimes, with fixed (stress) and random (spool) effects. The prior of the spool parameters is nonparametric, namely they are a sample from a normalized generalized gamma process, which encompasses the well‐known Dirichlet process. The nonparametric prior is assumed to robustify inferences to misspecification of the parametric prior. Here, this choice of likelihood and prior yields a new Bayesian model in reliability analysis. Via a Bayesian hierarchical approach, it is easy to analyze the reliability of the Kevlar fiber by predicting quantiles of the failure time when a new spool is selected at random from the population of spools. Moreover, for comparative purposes, we review the most interesting frequentist and Bayesian models analyzing this data set. Our credibility intervals of the quantiles of interest for a new random spool are narrower than those derived by previous Bayesian parametric literature, although the predictive goodness‐of‐fit performances are similar. Finally, as an original feature of our model, by means of the discreteness of the random‐effects distribution, we are able to cluster the spools into three different groups. Copyright © 2012 John Wiley & Sons, Ltd.     

A model for diffusion of mulucomponent information

The epidemic model of diffusion of news (or disease) is generalized to describe the diffusion of a multi‐component information. The multivaluedness of information in our model arises due to the large number (k) of constituent components or items of the information in question. When the different components of information are assumed to bear no hierarchy, the master equation of the model contains an intractably large number of variables (2 ^k ). The dynamics of the model, however, displays some simplifying features, one of which is the conservation of homogeneity of distribution of population over the different information vectors (in the sense defined in the text). The homogenized version of the model is found to be numerically tractable. The growth curves for large k continue to display sigmoid shapes, but with large ‘saturation times’. The dependence of ‘saturation time’ (i.e. the time required for spread of all the information in almost the entire population) on various parameters of the model, for uniform initial distributions, is numerically investigated. The ‘saturation time’ is found to vary inversely with the intensity of interaction (ß) and the size of population (N), as expected. An important numerical feature that emerges is that the ‘saturation time’ seems to be in linear proportion to the number of information items (k).