The classical theory of natural selection, as developed by Fisher, Haldane, and 'Wright, and their followers, is in a sense a statistical theory. By and large the classical theory assumes that the underlying environment in which evolution transpires is both constant and stable - the theory is in this sense deterministic. In reality, on the other hand, nature is almost always changing and unstable. We do not yet possess a complete theory of natural selection in stochastic environ ments. Perhaps it has been thought that such a theory is unimportant, or that it would be too difficult. Our own view is that the time is now ripe for the development of a probabilistic theory of natural selection. The present volume is an attempt to provide an elementary introduction to this probabilistic theory. Each author was asked to con tribute a simple, basic introduction to his or her specialty, including lively discussions and speculation. We hope that the book contributes further to the understanding of the roles of "Chance and Necessity" (Monod 1971) as integrated components of adaptation in nature.

In this dissertation, we introduce Stochastic CPPP (SCPPP), which is an extension of CPPP to domains with stochastic action effects. We show how SCPPP can be modeled as a Markov Decision Process (MDP) and how the VALUE-ITERATION algorithm can be adapted to solve it. This adaptation requires extending VALUE-ITERATION to support multiple agents and privacy. Then, we present two adaptions of the Real-Time Dynamic Programming (RTDP) algorithm, a popular algorithm for solving MDPs, designed to solve SCPPP problems. The first RTDP adaptation, called Distributed RTDP (DRTDP),yields identical behavior to applying RTDP in a centralized manner on the joint problem. To preserve privacy, DRTDP uses a message passing mechanism adopted from the Multi-Agent Forward Search (MAFS) algorithm. From the Abstract.

Evolutionary Algorithms (EAs) have grown into a mature field of research in optimization, and have proven to be effective and robust problem solvers for a broad range of static real-world optimization problems. Yet, since they are based on the principles of natural evolution, and since natural evolution is a dynamic process in a changing environment, EAs are also well suited to dynamic optimization problems. Evolutionary Optimization in Dynamic Environments is the first comprehensive work on the application of EAs to dynamic optimization problems. It provides an extensive survey on research in the area and shows how EAs can be successfully used to continuously and efficiently adapt a solution to a changing environment, find a good trade-off between solution quality and adaptation cost, find robust solutions whose quality is insensitive to changes in the environment, find flexible solutions which are not only good but that can be easily adapted when necessary. All four aspects are treated in this book, providing a holistic view on the challenges and opportunities when applying EAs to dynamic optimization problems. The comprehensive and up-to-date coverage of the subject, together with details of latest original research, makes Evolutionary Optimization in Dynamic Environments an invaluable resource for researchers and professionals who are dealing with dynamic and stochastic optimization problems, and who are interested in applying local search heuristics, such as evolutionary algorithms.

Populations living in natural environments experience fluctuations in environmental conditions that drive variability in demographic rates. This dissertation develops new and existing mathematical methods for studying environmental stochasticity and uses these tools to investigate the role of environmental stochasticity in driving observed population dynamics and plant life history evolution. In the first two chapters I develop new approaches to a classic method in population biology, the life table response experiment (LTRE). Whereas existing methods used time-averaged demographic rates and deterministic sensitivities to decompose observed differences in population growth rates, this new method allows estimation of the contributions to those differences made by variances in demographic rates as well as by mean rate values. I use this stochastic LTRE to show how differential variability in the vital rates of Anthyllis vulneraria (kidney vetch) contribute to differences in the population growth rates of nine populations growing in southwest Belgium; we also show how the effects of demographic rate variability depend on soil depth, where the greater moisture retention of deeper soils buffers populations against the otherwise negative effects of demographic variability. The second chapter provides a different approach to LTRE that uses an iterated two-factor decomposition of the small noise approximation of the stochastic population growth rate to quantify contributions to that growth rate made by: (i) mean vital rates, (ii) temporal variability in vital rates, (iii) elasticities of the population growth rate to individual vital rates, and (iv) correlations between vital rates across the study period. Contributions of elasticities tell us about differences in local selection pressures acting on distinct populations and contributions of correlations tell us about differences in the phenotypic tradeoffs associated with vital rates. I use this new method to show how these differences drive dynamics in two species: Anthyllis vulneraria (the same populations studied in the first chapter) and Cypripedium calceolus (lady's slipper orchid). In Anthyllis vulneraria, variability in large adult fertility and seedling survival made the largest contributions; there were also effects of differences in elasticities of large adult fertility and survival, as well as differences in the correlations between rapid growth and survival in seedlings (a survival cost of rapid early development), between large adult fertility and survival (a survival cost of reproduction) and between large adult fertility and seedling survival. In Cypripedium calceolus, population growth rates were driven most by differences in the elasticities to the probabilities of adult stasis vs. entering dormancy, as well as by differences in the variability and tradeoffs associated with adult dormancy; correlation played a role through differences in the survival payoff of dormancy vs. the complimentary fertility cost of dormancy in terms of lost opportunity for reproduction. The third and final chapter investigates the role of fire disturbance in driving the life histories and population-level dynamics of five woody plant species growing in the Brazilian cerrado, a savannah-forest mosaic in which woody vegetation cover is primarily mediated by fire disturbance. This study presents a set of diagnostics that use demographic responses to recurring disturbance to categorize species along a continuum of adaptation: on one end we find 'resistant' species that must weather disturbance in order to attain large sizes that are buffered against fire-induced mortality; on the other end we find 'resilient' species that are relatively indifferent to disturbance and harness transient opportunities afforded by early post-fire successional habitats in order to take advantage of increased nutrient availability and reduced competition. Each of these chapters uses stochastic demographic analysis to extend theory describing the dynamics of populations in variable environments; together, these studies present a variegated perspective on the role of environmental stochasticity that provides new methods and novel perspectives that should be useful in the study of population biology and life history evolution.

Adaptive agents and multi-agent systems is an emerging and exciting interdisciplinary area of research and development involving artificial intelligence, software engineering, and developmental biology, as well as cognitive and social science. This book presents 17 revised and carefully reviewed papers taken from two workshops on the topic as well as 2 invited papers by leading researchers in the area. The papers deal with various aspects of machine learning, adaptation, and evolution in the context of agent systems and autonomous agents.

Adaptive Control provides techniques for automatic, real-time adjustments in controller parameters with a view to achieving and/or maintaining a desirable level of system performance in the presence of unknown or variable process parameters. Many aspects of the field are dealt with in coherent and orderly fashion, starting with the problems posed by system uncertainties and moving on to the presentation of solutions and their practical significance. Within the general context of recent developments, the book looks at: • synthesis and analysis of parameter adaptation algorithms; • recursive plant-model identification in open and closed loop; • robust digital control for adaptive control; • direct and indirect adaptive control; and • practical aspects and applications. To reflect the importance of digital computers for the application of adaptive control techniques, discrete-time aspects are emphasized. To guide the reader, the book contains various applications of adaptive control techniques.

Adaptive Control (second edition) shows how a desired level of system performance can be maintained automatically and in real time, even when process or disturbance parameters are unknown and variable. It is a coherent exposition of the many aspects of this field, setting out the problems to be addressed and moving on to solutions, their practical significance and their application. Discrete-time aspects of adaptive control are emphasized to reflect the importance of digital computers in the application of the ideas presented. The second edition is thoroughly revised to throw light on recent developments in theory and applications with new chapters on: multimodel adaptive control with switching, direct and indirect adaptive regulation and adaptive feedforward disturbance compensation. Many algorithms are newly presented in MATLAB® m-file format to facilitate their employment in real systems. Classroom-tested slides for instructors to use in teaching this material are also now provided. All of this supplementary electronic material can be downloaded from fill in URL. The core material is also up-dated and re-edited to keep its perspective in line with modern ideas and more closely to associate algorithms with their applications giving the reader a solid grounding in: synthesis and analysis of parameter adaptation algorithms, recursive plant model identification in open and closed loop, robust digital control for adaptive control; • robust parameter adaptation algorithms, practical considerations and applications, including flexible transmission systems, active vibration control and broadband disturbance rejection and a supplementary introduction on hot dip galvanizing and a phosphate drying furnace. Control researchers and applied mathematicians will find Adaptive Control of significant and enduring interest and its use of example and application will appeal to practitioners working with unknown- and variable-parameter plant. Praise for the first edition: ...well written, interesting and easy to follow, so that it constitutes a valuable addition to the monographies in adaptive control for discrete-time linear systems... suitable (at least in part) for use in graduate courses in adaptive control.