Skip to main content

Yaneer Bar-Yam

New England Complex Systems Institute, President, Faculty Member

Followers

255

Following

19

Co-authors

16

Public Views

Interests

Uploads

Books by Yaneer Bar-Yam

Conflict and Complexity

by Phil Fellman, Ali Minai, and Yaneer Bar-Yam

This volume presents a complex systems approach to analyzing, modeling, understanding, and comba... more This volume presents a complex systems approach to analyzing, modeling, understanding, and combating terrorism and conflict, and is a unique and timely contribution to a topic of critical importance. Much of the effort in this area has used—and continues to use—classical methods based on intelligence, statistical and game theoretic modeling, and military operations. The need for other methods has become increasingly clear. Recognizing and conflict, at least in part, as a social phenomenon suggests that methods that have succeeded in analyzing other social systems may also work well in this case. This has led to the application of network modeling and analysis to terrorism and conflict. Other complex systems concepts such as chaotic dynamics, self- organization, emergent patterns, and fractals have also been applied, generating important insights. This book reviews and discusses these efforts.

Dynamics of Complex Systems Chapter 0

The study of complex systems in a unified framework has become recognized in recent years as a ne... more The study of complex systems in a unified framework has become recognized in recent years as a new scientific discipline, the ultimate of interdisciplinary fields. Breaking down the barriers between physics, chemistry and biology and the so-called soft sciences of psychology, sociology, economics, and anthropology, this text explores the universal physical and mathematical principles that govern the emergence of complex systems from simple components.

Dynamics of Complex Systems is the first text describing the modern unified study of complex systems. It is designed for upper-undergraduate/beginning graduate-level students, and covers a wide range of applications in a wide array of disciplines. A central goal of this text is to develop models and modeling techniques that are useful when applied to all complex systems. This is done by adopting both analytic tools, from statistical mechanics to stochastic dynamics, and computer simulation techniques, such as cellular automata and Monte Carlo. In four sets of paired, self-contained chapters, Yaneer Bar-Yam discusses complex systems in the context of neural networks, protein folding, living organisms, and finally, human civilization itself. He explores fundamental questions about the structure, dynamics, evolution, development and quantitative complexity that apply to all complex systems. In the first chapter, mathematical foundations such as iterative maps and chaos, probability theory and random walks, thermodynamics, information and computation theory, fractals and scaling, are reviewed to enable the text to be read by students and researchers with a variety of backgrounds.

Dynamics of Complex Systems Chapter 1

The study of complex systems in a unified framework has become recognized in recent years as a ne... more The study of complex systems in a unified framework has become recognized in recent years as a new scientific discipline, the ultimate of interdisciplinary fields. Breaking down the barriers between physics, chemistry and biology and the so-called soft sciences of psychology, sociology, economics, and anthropology, this text explores the universal physical and mathematical principles that govern the emergence of complex systems from simple components.

Dynamics of Complex Systems is the first text describing the modern unified study of complex systems. It is designed for upper-undergraduate/beginning graduate-level students, and covers a wide range of applications in a wide array of disciplines. A central goal of this text is to develop models and modeling techniques that are useful when applied to all complex systems. This is done by adopting both analytic tools, from statistical mechanics to stochastic dynamics, and computer simulation techniques, such as cellular automata and Monte Carlo. In four sets of paired, self-contained chapters, Yaneer Bar-Yam discusses complex systems in the context of neural networks, protein folding, living organisms, and finally, human civilization itself. He explores fundamental questions about the structure, dynamics, evolution, development and quantitative complexity that apply to all complex systems. In the first chapter, mathematical foundations such as iterative maps and chaos, probability theory and random walks, thermodynamics, information and computation theory, fractals and scaling, are reviewed to enable the text to be read by students and researchers with a variety of backgrounds.

Dynamics of Complex Systems Chapter 2

Motivated by the properties of biological neural networks, we introduce simple mathematical model... more Motivated by the properties of biological neural networks, we introduce simple mathematical models whose properties may be explored and related to aspects of human information processing. The attractor network embodies the properties of an associative content-addressable memory. Memories are imprinted and are accessed by presenting the network with part of their content. Properties of the network can be studied using a signal-to-noise analysis and simulations. The capacity of the attractor network for storage of memories is proportional to the number of neurons. The feedforward network acts as an input-output system formed out of several layers of neurons. Using prototypes that indicate the desired outputs for a set of possible inputs, the feedforward network is trained by minimizing a cost function which measures the output error. The resulting training algorithm is called back-propagation of error. In order to study the overall function of the brain, an understanding of sub-structure and the interactions between parts of the brain is necessary. Feedforward networks illustrate one way to build a network out of parts. A second model of interacting subnetworks is a subdivided attractor network. A subdivided attractor network stores more than just the imprinted patterns—it stores composite patterns formed out of parts of the imprinted patterns. If these are patterns that an organism might encounter , then this is an advantage. Features of human visual processing, language and motor control illustrate the relevance of composite patterns. Analysis and simulations of subdivided attractor networks reveal that partial subdivision can balance a decline in the storage capacity of imprinted patterns with the potential advantages of composite patterns. However, this balance only allows direct control over composite pattern stability when the number of subdivisions is no more than approximately seven, suggesting a connection to the 7 ± 2 rule of short-term memory.

Dynamics of Complex Systems Chapter 3

The training of a model network that has subdivisions requires a process that can train synapses ... more The training of a model network that has subdivisions requires a process that can train synapses within subdivisions and between subdivisions without subjecting either to overload. A natural solution to this problem involves taking the network " off-line, " so that a filtering of memories can occur when the network is dissociated. This is a possible model for the role of sleep in human information-processing that explains some of the unusual features of sleep and suggests new experiments that can be performed. Various features of human information-processing, including the learning of associations, pattern recognition, creativity, individuality and consciousness can be discussed within the context of neural network models. Efforts to describe and explain the higher information-processing tasks that human beings are capable of performing have always generated tension and concern. There has been a tendency to elevate these processes outside of the domain of the physical world, or to mystify them,through a characterization as infinite and incomprehensible. This tendency may arise from the desire to maintain a uniqueness of and importance to our own capabilities. We will adopt the contrary point of view that our capabilities are fundamentally comprehensible. However, it turns out that this does not diminish a quality of uniqueness and importance. If anything, it shows us how this importance arises. We begin to tackle the task of explaining aspects of human information-processing in this chapter. However, we will not conclude our discussion until we have described complexity in the context of human civilization in Chapter 9. We start in Section 3.1 by considering the training of subdivided net works and the role of sleep in human information-processing. This section is an essential sequel to the discussion in the last chapter that introduced subdivision. The problem is to develop a systematic approach to the training of subdivided networks .

Dynamics of Complex Systems Chapter 4

The simplest question about dynamics—how long does a process take?— becomes particularly relevant... more The simplest question about dynamics—how long does a process take?— becomes particularly relevant when the time may be so long that the process cannot happen at all. A fundamental problem associated with the dynamics of protein folding is understanding how a system of many interacting elements can reach a desired structure in a reasonable time. In this chapter, we discuss the parallel-processing idea for resolving this problem; kinetic pathways will be considered in the next chapter. Parallel processing and interdependence are at odds and must be balanced in the design of complex systems. We use finite-size Ising type models to explore the nature of interactions that can allow a system to relax in a time that grows less than exponentially in the size of the system. These models illustrate various ways to realize the parallel-processing idea. The simplest idealization of parallel processing is the case of completely independent spins. We discuss a two-spin model as a first example of how such a system relaxes. Various homogeneous models illustrate some of the properties that enable systems to relax in a time that grows no more than a power law in the system size. These include ideal parallel processing, and nucleation and growth of a stable state from a metastable state. The models also illustrate cases where exponential growth in the relaxation time can prevent systems from relaxing. Inhomogeneous models extend the range of possibilities for interaction architectures that still allow a reasonable relaxation time. Among these are space and time partitioning and preselected initial conditions. However, inhomogeneous long-range interactions generally lead to an exponential growth of relaxation time with system size.

Dynamics of Complex Systems Chapter 5

When kinetics limits the domain of phase space explored by a system, the scaling of the relaxatio... more When kinetics limits the domain of phase space explored by a system, the scaling of the relaxation time (N) may be smaller than exponential. Polymers in a liquid can be in an expanded or compact form. The transition between the two—poly-mer collapse—is a prototype of protein folding. Using simulations, we will explore possible origins of kinetic limitations in the phase space exploration of long polymers during collapse. Before we study collapse, we must understand the properties of polymers in their expanded state in good solvent. Simple arguments can tell us the scaling of polymer size, R (N) ~ N. The time scale of relaxation of a polymer from one confor-mation to another follows either Rouse (N) ~ N 2 +1 or Zimm (N) ~ N 3 scaling, depending on the assumptions used. Polymer simulations can be constructed in various forms. As long as they respect polymer connectivity and excluded volume, the behavior of long polymers is correctly reproduced. A two-space model where monomers alternate between spaces along the chain is a simple and convenient cellular automaton algorithm. During polymer collapse monomers bond and aggregate. Simulations of collapse and scaling arguments suggest that the aggregation occurs primarily at the ends of the polymer because of the greater flexibility of polymer-end motion. Thus the aggregates at the end appear to diffuse along the polymer contour accreting monomers and smaller aggregates until they meet in the middle. This results in an aggregation process that is systematically ordered by the kinetics. The end-dominated collapse-time scales linearly with polymer length, which is faster than the usual polymer relaxation. The orderly formation of bonds in end-dominated collapse also suggests that kinetics may constrain the possible monomer-monomer bonds that are formed and thus limit the domain of phase space that is explored in protein folding.

Dynamics of Complex Systems Chapter 6

The study of complex systems in a unified framework has become recognized in recent years as a ne... more The study of complex systems in a unified framework has become recognized in recent years as a new scientific discipline, the ultimate of interdisciplinary fields. Breaking down the barriers between physics, chemistry and biology and the so-called soft sciences of psychology, sociology, economics, and anthropology, this text explores the universal physical and mathematical principles that govern the emergence of complex systems from simple components.

Dynamics of Complex Systems is the first text describing the modern unified study of complex systems. It is designed for upper-undergraduate/beginning graduate-level students, and covers a wide range of applications in a wide array of disciplines. A central goal of this text is to develop models and modeling techniques that are useful when applied to all complex systems. This is done by adopting both analytic tools, from statistical mechanics to stochastic dynamics, and computer simulation techniques, such as cellular automata and Monte Carlo. In four sets of paired, self-contained chapters, Yaneer Bar-Yam discusses complex systems in the context of neural networks, protein folding, living organisms, and finally, human civilization itself. He explores fundamental questions about the structure, dynamics, evolution, development and quantitative complexity that apply to all complex systems. In the first chapter, mathematical foundations such as iterative maps and chaos, probability theory and random walks, thermodynamics, information and computation theory, fractals and scaling, are reviewed to enable the text to be read by students and researchers with a variety of backgrounds.

Dynamics of Complex Systems Chapter 7

Developmental biology strives to understand the sequence of events by which a single cell becomes... more Developmental biology strives to understand the sequence of events by which a single cell becomes a system of many differentiated interacting cells. This process involves placing different structures in particular locations and interconnecting them. To model differentiation we focus on the formation of color patterns on animal skins that have a variety of forms. Cellular automaton models show the relevance of local activation and long-range inhibition of pigment production to the formation of patterns. Chemical reaction-diffusion systems illustrate similar patterns using slow-and fast-diffusing species. Other elements of the tool kit for developmental processes include mechanisms for changes in cell structure, cell motion, timing and counting. Of particular interest are sequential steps (programs) that can form branching structures. Theoretical modeling can better complement phenomenological studies of biological systems if the different objectives of theory and experiment are recognized. The approach of developmental biology to the design of complex systems may be a useful framework for considering the design of complex artificial systems. Models of pattern formation may be better suited to discussions of global properties of the evolution of organisms than the models discussed in Chapter 6.

Dynamics of Complex Systems Chapter 8

Our ultimate objective is to consider the relationship of a human being to human civilization, wh... more Our ultimate objective is to consider the relationship of a human being to human civilization, where human civilization is considered as a complex system. We use this problem to motivate our study of the definition of complexity. The mathematical definition of the complexity of character strings follows from information theory. This theory is generalized by algorithmic complexity to allow all possible algorithms that can compress the strings. The complexity of a string is defined as the length of the shortest binary input to a universal Turing machine, such that the output is the string. The use of mappings from strings onto system states allows us to apply the concepts of algorithmic complexity to physical systems. However, the complexity of describing a microstate of the system is not really what we mean by system complexity. We define and study the complexity profile, which is the complexity of a system observed with a certain precision in space and time. We estimate the complexity of various systems, focusing on the complexity of a human being. Our final estimate is based upon a combination of the length of descriptions in human language, genetic information in DNA, and component counting.

Dynamics of Complex Systems Chapter 9

The study of complex systems in a unified framework has become recognized in recent years as a ne... more The study of complex systems in a unified framework has become recognized in recent years as a new scientific discipline, the ultimate of interdisciplinary fields. Breaking down the barriers between physics, chemistry and biology and the so-called soft sciences of psychology, sociology, economics, and anthropology, this text explores the universal physical and mathematical principles that govern the emergence of complex systems from simple components.

Dynamics of Complex Systems is the first text describing the modern unified study of complex systems. It is designed for upper-undergraduate/beginning graduate-level students, and covers a wide range of applications in a wide array of disciplines. A central goal of this text is to develop models and modeling techniques that are useful when applied to all complex systems. This is done by adopting both analytic tools, from statistical mechanics to stochastic dynamics, and computer simulation techniques, such as cellular automata and Monte Carlo. In four sets of paired, self-contained chapters, Yaneer Bar-Yam discusses complex systems in the context of neural networks, protein folding, living organisms, and finally, human civilization itself. He explores fundamental questions about the structure, dynamics, evolution, development and quantitative complexity that apply to all complex systems. In the first chapter, mathematical foundations such as iterative maps and chaos, probability theory and random walks, thermodynamics, information and computation theory, fractals and scaling, are reviewed to enable the text to be read by students and researchers with a variety of backgrounds.

Dynamics of Complex Systems Additional Readings

The following is a list of additional readings rather than a bibliography. The range of topics di... more The following is a list of additional readings rather than a bibliography. The range of topics discussed in this text does not allow for a comprehensive bibliography. Our focus is on the effort to develop concepts and methodologies that enable the study of complex systems in a unified manner. Nevertheless, this effort must be informed by many fields and their phenomenologies. The following list attempts to address this by providing accepted keywords for literature searches as provided by the Library of Congress. In addition to the keywords,a few references are provided with comments. Many of these texts were obtained from literature searches, and have been checked as relevant to the concepts we have been discussing. These references serve several purposes. First,they provide the student with an opportunity to pursue the phenomenology or theory in greater depth. Second, in a more specific domain, they provide a point of entry into the literature through a bibliography. Third, some references have an approach that is particularly compatible with the material presented in this text, or to the field of complex systems generally. This list,however, does not serve three conventional purposes. It does not serve to trace the historical origin of concepts presented, or to motivate them from phenomenological grounds, or to prove them using experimental observations. Any of these would be a worthwhile but equally challenging endeavor to the objective of demonstrating the unity of concepts, which is the motivating force behind this text. As is fitting for concepts that are to be a general underpinning of our understanding of complex systems, points made in this book appear in many contexts in the literature. A stronger statement may be made—the g enerality of the concepts presented in this text must imply that there are many ways to arrive at them,and many conclusions that may be drawn from them that can be compared with a large body of experimental literature. The effort in this text to draw conclusions from a very small set of assumptions is only a beginning in the effort to understand how widely applicable such concepts can be. In the few cases where we have made a greater effort to make contact with specific phenomenology and thus where support is necessary for material presented in the text (e.g. the discussion of sleep in Chapter 3), we have provided a few more specific references. There are a remarkable number of popular or semipopular books on various concepts in the study of complex systems. For a number of reasons these books have appeared instead of textbooks. They are of two types: books written by observers of the field, and books written by researchers presenting their ideas to a popular audience.

Dynamics of Complex Systems Index

The study of complex systems in a unified framework has become recognized in recent years as a ne... more The study of complex systems in a unified framework has become recognized in recent years as a new scientific discipline, the ultimate of interdisciplinary fields. Breaking down the barriers between physics, chemistry and biology and the so-called soft sciences of psychology, sociology, economics, and anthropology, this text explores the universal physical and mathematical principles that govern the emergence of complex systems from simple components.

Dynamics of Complex Systems is the first text describing the modern unified study of complex systems. It is designed for upper-undergraduate/beginning graduate-level students, and covers a wide range of applications in a wide array of disciplines. A central goal of this text is to develop models and modeling techniques that are useful when applied to all complex systems. This is done by adopting both analytic tools, from statistical mechanics to stochastic dynamics, and computer simulation techniques, such as cellular automata and Monte Carlo. In four sets of paired, self-contained chapters, Yaneer Bar-Yam discusses complex systems in the context of neural networks, protein folding, living organisms, and finally, human civilization itself. He explores fundamental questions about the structure, dynamics, evolution, development and quantitative complexity that apply to all complex systems. In the first chapter, mathematical foundations such as iterative maps and chaos, probability theory and random walks, thermodynamics, information and computation theory, fractals and scaling, are reviewed to enable the text to be read by students and researchers with a variety of backgrounds.

Papers by Yaneer Bar-Yam

The Precautionary Principle (with Application to the Genetic Modification of Organisms)

by R. Douady and Yaneer Bar-Yam

The precautionary principle (PP) states that if an action or policy has a suspected risk of causi... more The precautionary principle (PP) states that if an action or policy has a suspected risk of causing severe harm to the public domain (affecting general health or the environment globally), the action should not be taken in the absence of scientific near-certainty about its safety. Under these conditions, the burden of proof about absence of harm falls on those proposing an action, not those opposing it. PP is intended to deal with uncertainty and risk in cases where the absence of evidence and the incompleteness of scientific knowledge carries profound implications and in the presence of risks of "black swans", unforeseen and unforeseable events of extreme consequence.

Programed Death is Favored by Natural Selection in Spatial Systems

Physical Review Letters, Jan 12, 2015

Standard evolutionary theories of aging and mortality, implicitly based on mean-field assumptions... more Standard evolutionary theories of aging and mortality, implicitly based on mean-field assumptions, hold that programed mortality is untenable, as it opposes direct individual benefit. We show that in spatial models with local reproduction, programed deaths instead robustly result in long-term benefit to a lineage, by reducing local environmental resource depletion via spatiotemporal patterns causing feedback over many generations. Results are robust to model variations, implying that direct selection for shorter life span may be quite widespread in nature.

format_quoteThe model identifies spatial resource limitations as a mechanism that favors adaptive lifespan limitation over generations.format_quote

The Statistical Mechanics of Complex Product Development

In recent years, understanding the structure and function of complex networks has become the foun... more In recent years, understanding the structure and function of complex networks has become the foundation for explaining many different real-world complex biological, technological and informal social phenomena. Techniques from statistical physics have been successfully applied to the analysis of these networks, and have uncovered surprising statistical structural properties that have also been shown to have a major effect on their functionality, dynamics, robustness, and fragility. This paper examines, for the first time, the statistical properties of strategically important organizational networks - networks of people engaged in distributed product development (PD) - and discusses the significance of these properties in providing insight into ways of improving the strategic and operational decision-making of the organization. We show that the structure of information flow networks that are at the heart of large-scale product development efforts have properties that are similar to those displayed by other social, biological and technological networks. In this context, we identify novel properties that may be characteristic of other information-carrying networks. We further present a detailed model and analysis of PD dynamics on complex networks, and show how the underlying network topologies provide direct information about the characteristics of this dynamics. We believe that our new analysis methodology and empirical results are also relevant to other organizational information-carrying networks.

$To gain further insight regarding the rate of con- vergence to the fixed point of Equation (9), we solve it approximately as follows. For small values of a, Taylor’s expansion yields When the fraction of unresolved tasks at equilib- rium is very small (a* « 1), we can find a closed-form expression for a* by expanding f(a) in Equation (10) to the second order in @ and solving for a*:$

$Notes. The average number of incoming arcs connected to a node is (k,,) = 12.019. In (a), the time evolution of a(t) when the sensitivity and the internal completion rates of tasks are 8B = 0.061 and r = 0.75, respectively. In this case, B(K,,) <r, and the simulation run converges to the uniformly resolved state as predicted by theory. In (b), the sensitivity and the internal com- pletion rates are 6 = 0.064 and r = 0.75, respectively. In equilibrium, the average fraction of unresolved tasks between t = 65 and t = 100 (a station- ary regime) is a* = 0.087, which agrees reasonably well with the prediction a* = 0.090 given in Equation (12).$

Building Patterned Structures with Robot Swarms

International Joint Conference on Artificial Intelligence, 2005

We describe a system in which simple, identi- cal, autonomous robots assemble two-dimensional str... more We describe a system in which simple, identi- cal, autonomous robots assemble two-dimensional structures using prefabricated modules as build- ing blocks. Modules are capable of some infor- mation processing, enabling them to share long- range structural information and communicate it to robots. This communication allows arbitrary solid structures to be rapidly built using a few x ed, local robot behaviors.

Relationship between Measures of Fitness and Time Scale in Evolution

The notion of fitness is central in evolutionary biology. We use a simple spatially extended pred... more The notion of fitness is central in evolutionary biology. We use a simple spatially extended predator-prey or host-pathogen model to show a generic case where the average number of offspring of an individual as a measure of fitness fails to characterize the evolutionary dynamics. Mutants with high initial reproduction ratios have lineages that eventually go extinct due to local overexploitation. We propose
general quantitative measures of fitness that reflect the importance of time scale in evolutionary processes.

Information Flow Structure in Large-Scale Product Development Organizational Networks

Smart Business Networks, 2004

This paper analyzes the statistical properties of real-world networks of people engaged in produc... more This paper analyzes the statistical properties of real-world networks of people engaged in product development (PD) activities. We show that complex PD networks display similar statistical patterns to other real-world complex social, information, biological and technological networks. The paper lays out the foundations for understanding the properties of other intra- and inter-organizational networks that are realized by specific network architectures.

Untangling the Information Web of Complex System Design

Symposium on Complex Systems Engineering

Understanding the structure and function of complex networks has recently become the foundation f... more Understanding the structure and function of complex networks has recently become the foundation for explaining many different real-world complex biological, technological and informal social phenomena. The analysis of these networks has uncovered surprising statistical structural properties that have also been shown to have a major effect on their functionality, dynamics, robustness, and fragility. This paper examines the statistical properties of

format_quoteLarger percentages of heavy information consumers/generators exacerbate 'solve-and-rework' modes in PD networks.format_quote