NK simulation modeling

In high-tech industrial clusters as the aerospace most collaborations for innovations are highly knowledgespecific and form a (relatively dense) knowledge network. With reference to the case of the aerospace industrial cluster of the Lazio Region, in this work we study the network dynamics of its core organizations (firms, universities, and research centers). By applying the methodology of NK simulation modeling, we explore what happens lacking the initial contribution of knowledge provided by universities and research centers. Further, we investigate the effects of the introduction of minimum requisites in terms of knowledge heterogeneity and knowledge amount. We show that, within a general favorable condition of activation rules, trajectories are quite short and knowledge dynamics is sensitive to the requisite of knowledge heterogeneity, when it is set-up around half of its potential range. We will conclude that, despite some interesting results like the ones we have found and many others that could be discovered, this methodology has substantial failures mostly due to its requisites of computational burden and topological and behavioral invariance, which makes it hardly applicable and scarcely informative into empirical analysis of phenomena within economics and management sciences.

Notwithstanding the central place covered inside organization science and the economic theory of the firm, organization design theory still lacks sound building blocks concerning the effects that some fundamental variables have on workgroup performance. In this chapter a contribution to fill in this gap is given with reference to the relationships between connection modes and performance. In particular, through an agent-based simulation model a number of experiments have been done respect to the moderating role played by group size and task complexity. Results confirm current (but not really scientific) knowledge, and bring forth our understanding of these fundamental (and mostly nonlinear) relationships. Among the main results, it can be underlined that the best combinations between connection modes, task complexity, and workgroup size occur when complex tasks are connected by mutual adaptation and run by a small number of agents, or when less complex tasks are connected by paralle...

In this chapter we test the hypothesis that uneven links distributions and uneven absorptive capacity between an industrial cluster members provide some kind of competitive advantages. Through an agent-based model has been built and calibrated on real data taken from an aerospace industrial cluster, that hypothesis is contrasted against the normal, the uniform, and the U-shaped distribution. The focus of the model is on knowledge variables, agents' learning capacities, and structural variables, like firms size and proximity. Physical production is not considered, excepted for its degree of complexity, which determines also the degree of knowledge complexity. This work shows that, actually, the best performance in terms of cluster knowledge creation, growth and diffusion is obtained when firms connectedness and absorptive capacity are distributed in a scale-free way. More generally, the more unbalanced are these two variables (especially absorptive capacity), the better is knowle...

In high-tech industrial clusters as the aerospace most collaborations for innovations are highly knowledge-specific and form a (relatively dense) knowledge network. With reference to the case of the aerospace industrial cluster of the Lazio Region, in this work we study the network dynamics of its core organizations (firms, universities, and research centers). By applying the methodology of NK simulation modeling, we explore what happens lacking the initial contribution of knowledge provided by universities and research centers. Further, we investigate the effects of the introduction of minimum requisites in terms of knowledge heterogeneity and knowledge amount. We show that, within a general favorable condition of activation rules, trajectories are quite short and knowledge dynamics is sensitive to the requisite of knowledge heterogeneity, when it is set-up around half of its potential range. We will conclude that, despite some interesting results like the ones we have found and ma...

Do trading countries also collaborate in R&D? This is the question that, facing with a number of methodological problems, here it is dealt with. Studying and comparing the international trade network and the R&D collaboration network of European countries in the aerospace sector, social network analysis offers a wide spectrum of methods and criteria either to make them comparable or to evaluate its similarity. International trade is a 1-mode directed and valued network, while the EU-subsidized R&D collaboration is an affiliation (2-mode) undirected and unvalued network, and the elementary units of this latter are organizations and not countries. Therefore, to the aim to make these two networks comparable, this paper shows and discusses a number of methodological problems and solutions offered to solve them, and provides a multi-faceted comparison in terms of various statistical and topological indicators. A comparative analysis of the two networks structures is made at aggregate and...

This material is brought to you by the International Conference on Information Systems (ICIS) at AIS Electronic Library (AISeL). It has been accepted for inclusion in ICIS 2010 Proceedings by an authorized administrator of AIS Electronic Library (AISeL). For more information, ...

An ultimate goal of a molecular biologist is to use genetic data to reveal fundamental cellular processes, and their impact on complex organisms. In order to achieve this goal one has to study how complex systems of several genes and proteins function and interact.

An ultimate goal of a molecular biologist is to use genetic data to reveal fundamental cellular processes, and their impact on complex organisms. In order to achieve this goal one has to study how complex systems of several genes and proteins function and interact.

for supervising and guiding my work. Sarabajaya Kumar from the LSE Teaching and Learning Centre for giving me the support for completion. AVINA Stiftung for the resources to make this thesis more than an academic work. Stephan Schmidheiny for his courage to 'walk the talk' of Sustainability and his generosity to society. Finally, I am deeply grateful to Peter Fuchs for sharing his wisdon on social change and encouranging me to follow my ideas.

There are three main aims of this paper. 1- I explain reasons why I await life to lie significantly deeper in chaos than Kauffman approach does, however still in boundary area near `the edge of chaos and order'. The role of negative feedbacks in stability of living objects is main of those reasons. In Kauffman's approach regulation using negative feedbacks is not considered sufficiently, e.g. in gene regulatory model based on Boolean networks, which indicates therefore not proper source of stability. Large damage avalanche is available only in chaotic phase. It models death in all living objects necessary for Darwinian elimination. It is the first step of my approach leading to structural tendencies which are effects of adaptive evolution of dynamic complex (maturely chaotic) networks. 2- Introduction of s>=2 equally probable variants of signal (state of node in Kauffman network) as interpretively based new statistical mechanism (RSN) instead of the bias p - probability o...

Control over suppliers is a core issue for a buying firm. Despite the amount of research on the topic, how much of its supply network (i.e., scope of control) a buying firm should control, so as to increase supply network performance, is a question that has not received adequate research attention. The study addresses this research question by considering the supply network as a complex adaptive system and developing an empirically informed agent-based simulation model using the NK fitness landscape framework, to examine how varying levels of scope of control influence supply network performance. We also investigate the direct effect of two supply network complexity dimensions (i.e., number of firms and level of supply interactions) on supply network performance and the moderating effect played by the scope of control. Results show that the relationship between scope of control and supply network performance follows an inverted-U shape. Furthermore, we find that the complexity dimensions negatively affect supply network performance with the performance decrease depending on the scope of control. Based on these findings, we formulate different control strategies to mitigate the negative influence of complexity.

In practice, product design decisions in a buyer-supplier setting can be either centralized or decentralized. Yet, it is unclear in what circumstances either of these settings can result in higher new product performance. In this study, we show that the choice of the level of centralization of product design decisions should be made contingent to the product complexity and behavioral factors of project managers involved in the decisions. In particular, we examine the influence of the cognitive ability, resistance to change, and cognitive effort of the project managers on the choice of the level of centralization of product design decisions. A methodology coming from complexity science is adopted and a simulation analysis is carried out. Results confirm that product complexity affects the relationship between the level of central-ization of the product design decisions and product performance. Furthermore, we show that the benefit of centralized product design decisions is influenced by the level of cognitive effort, cognitive ability, and resistance to change of project managers. The analysis is performed in two different contexts characterized by varying levels of trust between the buyer and supplier. We discuss the implications of our findings for theory and practice. Index Terms-Human behavior, new product development (NPD) process, organization design, project success factors, simulation , social psychology, supply chain integration.

This is a simulation-based paper that aims to understand the impact of adaptive long-jumps on innovation landscapes. In the context of a given technological space, the set of innovative outcomes is conceived as a landscape, with the various possible outcomes seen as dotting different fitness peaks. Organizations, in their pursuit of innovation, undertake adaptive walks over this landscape. It has been established that progress in scientific knowledge enables organizations to 'look beyond' local optima. We model this behaviour through the provision of long-jumps. We study the effect of organizations undertaking the best possible combination of long-jumps and 1-mutant walks towards their eventual landing forms. We expect to see 'basins of attraction' forming around global optima which lead to realization of concentrated structure. In effect, we find that this may lead to a reversal of the Schumpeterian hypothesis by showing that innovation strategies can influence market concentration.

Automatic techniques for the design of artificial computational systems, such as control programs for robots, are currently achieving increasing attention within the AI community. A prominent case is the design of artificial neural network systems by means of search techniques, such as genetic algorithms. Frequently, the search calibrates not only the system parameters, but also its structure. This procedure has the advantage of reducing the bias introduced by the designer and makes it possible to explore new, innovative solutions. The drawback, though, is that the analysis of the resulting system might be extremely difficult and limited to few coarse-grained characteristics. In this paper, we consider the case of robots controlled by Boolean networks that are automatically designed by means of a training process based on local search. We propose to analyse these systems by a method that detects mesolevel dynamical structures. These structures are emerging patterns composed of elements that behave in a coherent way and loosely interact with the rest of the system. In general, this method can be used to detect functional clusters and emerging structures in nonlinear discrete dynamical systems. It is based on an extension of the notion of cluster index, which has been previously proposed by Edelman and Tononi to analyse biological neural systems. Our results show that our approach makes it possible to identify the computational core of a Boolean network which controls a robot.

In high-tech industrial clusters as the aerospace most collaborations for innovations are highly knowledgespecific
and form a (relatively dense) knowledge network. With reference to the case of the aerospace
industrial cluster of the Lazio Region, in this work we study the network dynamics of its core organizations
(firms, universities, and research centers). By applying the methodology of NK simulation modeling,
we explore what happens lacking the initial contribution of knowledge provided by universities and
research centers. Further, we investigate the effects of the introduction of minimum requisites in terms
of knowledge heterogeneity and knowledge amount. We show that, within a general favorable condition
of activation rules, trajectories are quite short and knowledge dynamics is sensitive to the requisite of
knowledge heterogeneity, when it is set-up around half of its potential range. We will conclude that,
despite some interesting results like the ones we have found and many others that could be discovered,
this methodology has substantial failures mostly due to its requisites of computational burden and topological
and behavioral invariance, which makes it hardly applicable and scarcely informative into
empirical analysis of phenomena within economics and management sciences.

Sociology and other social sciences have employed network analysis earlier than management and
organization sciences, and much earlier than economics, which has been the last one to systematically
adopt it. Nevertheless, the development of network economics during last 15 years has been massive,
alongside three main research streams: strategic formation network modeling, (mostly descriptive)
analysis of real economic networks, and optimization methods of economic networks. The main reason
why this enthusiastic and rapidly diffused interest of economists came so late is that the most essential
network properties, like externalities, endogenous change processes, and nonlinear propagation
processes, definitely prevent the possibility to build a general – and indeed even partial – competitive
equilibrium theory. For this paradigm has dominated economics in the last century, this incompatibility
operated as a hard brake, and presented network analysis as an inappropriate epistemology. Further,
being intrinsically (and often, until recent times, also radically) structuralist, social network analysis
was also antithetic to radical methodological individualism, which was – and still is – economics dominant
methodology. Though culturally and scientifically influenced by economists in some fields, like
finance, banking and industry studies, scholars in management and organization sciences were free
from “neoclassical economics chains”, and therefore more ready and open to adopt the methodology
and epistemology of social network analysis. The main and early field through which its methods were
channeled was the sociology of organizations, and in particular group structure and communication,
because this is a research area largely overlapped between sociology and management studies. Currently,
network analysis is becoming more and more diffused within management and organization sciences.
Mostly descriptive until 15 years ago, all the fields of social network analysis have a great opportunity
of enriching and developing its methods of investigation through statistical network modeling, which
offers the possibility to develop, respectively, network formation and network dynamics models. They are
a good compromise between the much more powerful agent-based simulation models and the usually
descriptive (or poorly analytical) methods.

Launched and developed primarily by Kauffman from the end of sixties, NK simulation modelling candidates for capturing networks dynamics. Grounded in reference to biological networks, it has aroused a grate and durable interest in economics and management sciences too. This methodology is split into a version focused on studying proper Boolean networks dynamics, whose trajectories are substantially
conditioned by Boolean functions, and a (much more frequented) version focused on systems co-evolutionary
paths driven by the search for optimizing its fitness value. Besides the unquestionable value of Kauffman’s work for the theoretical implications on evolutionary biology and the strong interest for economics and management sciences, in this chapter failures and limitations of both NK modelling versions
are discussed. In particular, it is shown that as applications try to be more realistic, this modeling becomes hardly treatable from a computational point of view. On the other hand, it is underlined that, especially the fitness landscape version, NK simulation modelling is very useful to show general aspects
of system’s dynamics, and the impossibility to find general optima (excepted for very special and unrealistic cases). This result sounds a sharp criticism to general economic equilibrium, and it is perfectly consistent with Simon’s contributions.

This study presents a new model of search on a rugged landscape, which employs modeling techniques from fractal geometry rather than the now-familiar NK-modeling technique. In our simulations, firms search locally in a two-dimensional fitness landscape, choosing moves in a way that responds both to local payoff considerations and to a more global sense of opportunity represented by a firm-specific preferred direction. The latter concept provides a simple device for introducing cognitive or motivational considerations into the formal account of search behavior, alongside payoff considerations. After describing the objectives and the structure of the model, we report a first experiment that explores how the ruggedness of the landscape affects the interplay of local payoff and cognitive considerations (preferred direction) in search. We show that an intermediate search strategy, combining the guidance of local search with a moderate level of nonlocal obsession, is distinctly advantageous in searching a rugged landscape. We also explore the effects of other considerations, including the objective validity of the preferred direction and the degree of dispersion of firm strategies. We conclude by noting available features of the model that are not exercised in this experiment. Given the inherent flexibility of the model, the range of questions that might potentially be explored is extremely large.

A research program focusing on understanding and modelling the evolution of industrial networks is described which draws on recent theoretical and modeling advances in the area of artificial life and complex adaptive systems. The creation and evolution of the structure of industrial networks is seen to emerge through the on-going micro processes of action and interaction among and within firms. In this paper we focus attention on a particular class of models develop by Stuart Kauffman called NK Models, which permit key elements of industrial structures to be modelled effectively. Using this approach we develop a model of the interaction between dyadic relations in industrial networks. Future developments of this type of research are discussed as well as its implications for understanding, predicting and coping

The question of group performance is one of the most central in organization science. However until now the studies have been lacking theoretical clarity, formal demonstrations, and empirical tests. Focusing on technological interdependence, we will identify and define the fundamental types of systemic couplings: parallel, sequential and reciprocal (by feedback). Then we will precisely discuss and then simulate its characteristics and effects through an agent-based model. We measure and order the relational complexity, that is, the degree of complexity of each systemic coupling (technological interdependence between tasks). Group performance will be measured by means of two main indexes: effectiveness, which refers to the percentage of completed tasks respect to those potentially executable; and efficiency, which refers to the resources employed for getting the completed tasks. We analyze the variation of group performance depending on relational complexity. The most general and important conclusions are that: i) listing in order of increasing degree of complexity, parallel is followed by sequential and then by reciprocal interdependence; ii) more complex interdependencies do require more complex norms. This latter conclusion is an exemplar confirmation of the law of requisite variety.

ABSTRACT Information systems development (ISD) is essentially a search process by which the ISD team seeks to find an optimal system configuration among numerous feasible configurations that produces the greatest business value. As information systems are embodiments of application domain knowledge and technical knowledge, ISD requires both in order to be effective.

Abstract Supply chains, or supply networks (SNs), exist in a multitude of different topologies, yet little is known concerning how such topologies grow, evolve, and adapt over time. To study this complex phenomenon, we begin by identifying some primary topological structures that SNs may form. Then, to investigate the evolution of such structures, a theory-based framework is developed that combines aspects of complex adaptive systems theory, industrial growth theory, network theory, market structure, and game theory.

ABSTRACT Information systems development (ISD) is essentially a search process by which the ISD team seeks to find an optimal system configuration among numerous feasible configurations that produces the greatest business value. As information systems are embodiments of application domain knowledge and technical knowledge, ISD requires both in order to be effective.