Neural Approximations for Optimal Control and Decision provides a comprehensive methodology for the approximate solution of functional optimization problems using neural networks and other nonlinear approximators where the use of traditional optimal control tools is prohibited by complicating factors like non-Gaussian noise, strong nonlinearities, large dimension of state and control vectors, etc.

Features of the text include:

. a general functional optimization framework;

. thorough illustration of recent theoretical insights into the approximate solutions of complex functional optimization problems;

. comparison of classical and neural-network based methods of approximate solution;

. bounds to the errors of approximate solutions;

. solution algorithms for optimal control and decision in deterministic or stochastic environments with perfect or imperfect state measurements over a finite or infinite time horizon and with one decision maker or several;

. applications of current interest: routing in communications networks, traffic control, water resource management, etc.; and

. numerous, numerically detailed examples.

The authors' diverse backgrounds in systems and control theory, approximation theory, machine learning, and operations research lend the book a range of expertise and subject matter appealing to academics and graduate students in any of those disciplines together with computer science and other areas of engineering.

Autorentext

Riccardo Zoppoli received the "Laurea" degree in electronic engineering from the University of Genoa, Italy, in 1965, and the title of "Libero Docente" in 1971. From 1968 to 1979, he was an Associate Professor of Control Theory at the University of Genoa. Since 1980, he has been Full Professor of Control Theory and Operations Research at the same University. In 1972, was a NATO Visiting Scholar at the University of California at Los Angeles. From 1979 to 1985, he was responsible for the special project "Distributed Process Control Systems" supported by the National Research Council of Italy. He is the author of more than 100 papers scientific papers in the area of systems theory and operations research. His more recent research interests include large-scale systems, team theory, optimal control and filtering, and neural approximations. Prof. Zoppoli has served on various national and international and academic committees and as member of several boards of technical boards. In 2017, he received the title of "Professor Emeritus" from University of Genova, Italy.

Marcello Sanguineti  was born in Chiavari (Genoa), Italy. He received the ``Laurea'' (MSc) degree cum laude in Electronic Engineering and the PhD degree in Electronic Engineering and Computer Science, both from the University of Genova (Italy). He is currently Associate Professor of Operations Research at the University of Genova, with National Scientific Qualification as Full Professor, where he teaches "Operations Research", "Optimization Techniques" and "Strategic Choices: Games and Team Optimization". He is also Research Associate at CNR-INM (Institute for Marine Engineering) and Guest Scholar at IMT - School for Advances Studies (Lucca, Italy). Previously he was Research Associate at CNR-ISSIA (Institute of Intelligent Systems for Automation). He was member of the Scientific Boards of several PhD programs at the University of Genova, currently the PhD program in Computer Science and Systems Engineering. Marcello Sanguineti spent several periods as visiting scientist and lecturer at the Institute of Computer Science of the Czech Academy of Sciences and at the Department of Management of Bar-Ilan University (Tel Aviv, Israel). He was project referee for the Czech Academy of Sciences, for the Georgia National Science Foundation (GNSF), and for the Italian Ministry of Education, University, and Research (MIUR). He authored or coauthored more than 200 research papers in archival journals, book chapters, and international conference proceedings. He was member of the Program Committees of several conferences, member of the Organizing Committee of the 2007 Conference of the Italian Association of Operations Research, and Chair of the Organizing Committee of the 2008 International Conference on Mathematical Problems in Engineering, Aerospace, and Sciences. He is currently member of the Organizing Committee of the 2019 International Conference on Optimization and Decision Science. Marcello Sanguineti was invited speaker and keynote speaker at several international and Italian conferences. He coordinated international and Italian research projects on approximate solution of optimization problems and nonlinear approximation by neural networks. He was Principal Investigator of a NATO Collaborative Linkage Grant and member of several European projects (H2020 and 7th Framework Program) and other international projects. He was Guest Editors of the journals Computers and Operations Research and Computational Management Science. He is currently serving as an Associate Editor of the journals Neurocomputing and Neural Processing Letters and from 2006 to 2012 he was Associate Editor of the IEEE Trans. on Neural Networks. His more recent research interests are machine learning, neural networks for optimization, infinite-dimensional programming, network optimization, game theory, and affective computing.

Giorgio Gnecco was born in Genoa, Italy. From 2017, he is a Tenure-Track Assistant Professor ("Ricercatore Tipo B") in Operations Research at the IMT - School for Advanced Studies, Lucca, working in the Analysis of compleX Economic Systems (AXES) research unit, with National Scientific Qualification as Full Professor in Operations Research from 2018. From 2013 to 2017, he was a Fixed Term Assistant Professor ("Ricercatore Tipo A") in Control Systems at IMT Lucca, working in the Dynamical Systems, Control and Optimization (DYSCO) research unit. From 2016, he is Director of the local Research Unit of INdAM (National Institute for Advanced Mathematics) at IMT - Lucca. He has been an Associate Editor for IEEE Transactions on Neural Networks and Learning Systems (in the first quartile in 4 different subject categories according to the ISI Impact Factor 2017) since January 2013. In March 2018, he taught the graduate course "Optimization...

Klappentext
Neural Approximations for Optimal Control and Decision provides a comprehensive methodology for the approximate solution of functional optimization problems using neural networks and other nonlinear approximators where the use of traditional optimal control tools is prohibited by complicating factors like non-Gaussian noise, strong nonlinearities, large dimension of state and control vectors, etc.

Features of the text include:

• a general functional optimization framework;

• thorough illustration of recent theoretical insights into the approximate solutions of complex functional optimization problems;

• comparison of classical and neural-network based methods of approximate solution;

• bounds to the errors of approximate solutions;

• solution algorithms for optimal control and decision in deterministic or stochastic environments with perfect or imperfect state measurements over a finite or infinite time horizon and with one decision maker or several;

• applications of current interest: routing in communications networks, traffic control, water resource management, etc.; and

• numerous, numerically detailed examples.

The authors' diverse backgrounds in systems and control theory, approximation theory, machine learning, and operations research lend the book a range of expertise and subject matter appealing to academics and graduate students in any of those disciplines together with computer science and other areas of engineering.

Zusammenfassung

Many areas of science and technology require the soluti…