Studies in Computational Intelligence 854 Erik Cuevas Jorge Gálvez Omar Avalos Recent Metaheuristics Algorithms for Parameter Identification Studies in Computational Intelligence Volume 854 Series Editor Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland The series “Studies in Computational Intelligence” (SCI) publishes new develop- mentsandadvancesinthevariousareasofcomputationalintelligence—quicklyand with a high quality. The intent is to cover the theory, applications, and design methods of computational intelligence, as embedded in the fields of engineering, computer science, physics and life sciences, as well as the methodologies behind them. The series contains monographs, lecture notes and edited volumes in computational intelligence spanning the areas of neural networks, connectionist systems, genetic algorithms, evolutionary computation, artificial intelligence, cellular automata, self-organizing systems, soft computing, fuzzy systems, and hybrid intelligent systems. Of particular value to both the contributors and the readership are the short publication timeframe and the world-wide distribution, which enable both wide and rapid dissemination of research output. The books of this series are submitted to indexing to Web of Science, EI-Compendex, DBLP, SCOPUS, Google Scholar and Springerlink. More information about this series at http://www.springer.com/series/7092 á Erik Cuevas Jorge G lvez (cid:129) (cid:129) Omar Avalos Recent Metaheuristics Algorithms for Parameter fi Identi cation 123 ErikCuevas Jorge Gálvez CUCEI CUCEI Universidad deGuadalajara Universidad deGuadalajara Guadalajara, Mexico Guadalajara, Mexico OmarAvalos CUCEI Universidad deGuadalajara Guadalajara, Mexico ISSN 1860-949X ISSN 1860-9503 (electronic) Studies in Computational Intelligence ISBN978-3-030-28916-4 ISBN978-3-030-28917-1 (eBook) https://doi.org/10.1007/978-3-030-28917-1 ©SpringerNatureSwitzerlandAG2020 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar methodologynowknownorhereafterdeveloped. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained hereinorforanyerrorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregard tojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface Sinceancienttimes,humanshaveusedrulesofthumbandstrategiesextractedfrom other domains to solve several conflicting problems. Much of these methods have been adopted in different areas such as economy, engineering. These problem-solvingtechniquesarereferredundertheconceptofheuristics.Ingeneral, heuristics are specially considered in those situations where there is insufficient or incomplete information about the problem to be solved. As several problems pre- sentthischaracteristic,heuristics-basedtechniquesbecamequitepopularinthelast years. On the other hand, metaheuristics imply a high-level heuristics in which the problem-solving strategy is more general and adaptable to multiple contexts. Metaheuristic methods perform better than simple heuristics, since their mecha- nisms are guided for information (or knowledge) embedded within the problem-solving process. Metaheuristic methods use as inspiration our scientific understanding of biological, natural, or social systems, which at some level of abstraction can be represented as optimization processes. They intend to serve as general-purpose easy-to-use optimization techniques capable of reaching globally optimal or at least nearly optimal solutions. In their operation, searcher agents emulateagroupofbiologicalorsocialentitieswhichinteractwitheachotherbased on specialized operators that model a determined biological or social behavior. Theseoperatorsareappliedtoapopulationofcandidatesolutions(individuals)that are evaluated with respect to an objective function. Thus, in the optimization processindividualpositionsaresuccessivelyattractedtotheoptimalsolutionofthe system to be solved. There exist several features that clearly appear in most of the metaheuristic approaches, such as the use of diversification to force the exploration of regions of the search space, rarely visited until now, and the use of intensification or exploitation,toinvestigatethoroughlysomepromisingregions.Anotherinteresting feature is the use of memory to store the best solutions encountered. For these reasons, metaheuristics methods quickly became popular among researchers to solve from simple to complex optimization problems in different areas. v vi Preface Most of the problems in science, engineering, economics, and life can be translatedasanoptimizationorasearchproblem.Accordingtotheircharacteristics, some problems can be simple that can be solved by traditional optimization methods based on mathematical analysis. However, most of the problems of practical importance such as system identification, parameter estimation, energy systems represent conflicting scenarios so that they are very hard to be solved by using traditional approaches. Under such circumstances, metaheuristic algorithms have emerged as the best alternative to solve this kind of complex formulations. Therefore, metaheuristics techniques have consolidated as a very active research subjectinthelasttenyears.Duringthistime,variousnewmetaheuristicapproaches have been introduced. They have been experimentally examined on a set of arti- ficial benchmark problems and in a large number of practical applications. Although metaheuristic methods represent one of the most exploited research paradigms in computational intelligence, there are a large number of open chal- lenges in the area of metaheuristics. They range from premature convergence, inability to maintain population diversity, and the combination of metaheuristics withotheralgorithmicschemes,towardextendingtheavailabletechniquestotackle ever more difficult problems. Amongtheengineeringproblems,identificationsystemsofprocesses,whichare nonlinear in nature, represent a challenging formulation. In general, identification systems refer to methods which allow to estimate the parameters that mathemati- cally model a certain process. From an optimization perspective, identification systems are considered extremely complex due to their nonlinearity, discontinuity, andhighmultimodality.Thesecharacteristicsmakedifficulttosolvethembyusing traditional optimization techniques. In the last years, researchers, engineers, and practitioners in identification systems and modeling have faced problems of increasingcomplexity.Theseproblemscanbe stated asoptimizationformulations. Under these circumstances, an objective function is defined to evaluate the quality of each candidate solution composed of the problem parameters. Then, an opti- mization method is used to find the best solution that minimizes/maximizes the objective function. Numerousbookshavebeenpublishedtackinginaccountanyofthemostwidely known metaheuristic methods, namely simulated annealing, tabu search, evolu- tionary algorithms, ant colony algorithms, particle swarm optimization or differ- ential evolution, but attempts to consider the discussion of new alternative approachesarealwaysscarce.Initialmetaheuristicschemesmaintainintheirdesign several limitations such as premature convergence and inability to maintain pop- ulation diversity. Recent metaheuristic methods have addressed these difficulties providing in general better results. Many of these novel metaheuristic approaches have also been lately introduced. In general, they propose new models and inno- vative algorithm combinations for producing an adequate exploration and exploitationoflargesearchspacesconsideringasignificantnumberofdimensions. Most of the new metaheuristic algorithms present promising results. Nevertheless, they arestill intheirinitial stage.To growandattaintheircompletepotential,new metaheuristicmethodsmustbeappliedinagreatvarietyofproblemsandcontexts, Preface vii so that they do not only perform well in their reported sets of optimization prob- lems, but also in new complex formulations. The only way to accomplish this is making possible the transmission and presentation of these methods in different technical areas as optimization tools. In general, once a researcher, engineer, or practitioner recognizes a problem as a particular instance of a more generic class, he/she can select one of the different metaheuristic algorithms that guarantee an expected optimization performance. Unfortunately, the set of options are concen- tratedinalgorithmswhosepopularityandhighproliferationarebetterthanthenew developments. The excessive publication ofdevelopments based on thesimple modificationof popularmetaheuristicmethodspresentsanimportantdisadvantage:Theyavoidthe opportunity to discover new techniques and procedures which can be useful to solve problems formulatedbytheacademic andindustrial communities.Inthelast years, several promising metaheuristic methods that consider very interesting concepts and operators have been introduced. However, they seem to have been completely overlooked in the literature, in favor of the idea of modifying, hybridizing, or restructuring popular metaheuristic approaches. The first goal of this book is to present advances that discuss new alternative metaheuristic developments which have proved to be effective in their application to several complex problems. The book considers different new metaheuristic methods and their practical applications. This structure is important to us, because we recognize this methodology as the best way to assist researchers, lecturers, engineers, and practitioners in the solution of their own optimization problems. The second goal of this book is to bridge the gap between recent metaheuristic techniques with interesting identification system methods that profit on the con- venient properties of metaheuristic schemes. To do this, at each chapter we endeavor to explain basic ideas of the proposed applications in ways that can be understood by readers who may not possess the necessary backgrounds on either ofthefields.Therefore,identificationsystemsandenergypractitionerswhoarenot researchers in metaheuristics will appreciate that the techniques discussed are beyond simple theoretical tools since they have been adapted to solve significant problems that commonly arise on such areas. On the other hand, members of the metaheuristic community can learn the way in which system identification and energy problems can be translated into optimization tasks. This book has been structured so that each chapter can be read independently from the others. Chapter 1 describes the main characteristics and properties of metaheuristic methods. This chapter concentrates on elementary concepts of metaheuristic. Readers that are familiar with metaheuristic algorithms may wish to skip this chapter. In Chap. 2, an algorithm for the optimal parameter identification of induction motors is presented. In the identification, the parameter estimation process is transformed into a multidimensional optimization problem where the internal parameters of the induction motor are considered as decision variables. Under this approach,thecomplexityoftheoptimizationproblemtendstoproducemultimodal error surfaces for which their cost functions are significantly difficult to minimize. viii Preface To determine the parameters, the presented scheme uses a recent metaheuristic method called the gravitational search algorithm (GSA). Different to the most of existentevolutionaryalgorithms,GSApresentsabetterperformanceinmultimodal problems,avoidingcritical flawssuchastheprematureconvergencetosuboptimal solutions. Numerical simulations have been conducted on several models to show the effectiveness of the proposed scheme. Chapter 3 presents an improved version of the crow search algorithm (CSA)methodtosolvecomplexoptimizationproblemsofenergy.Intheimproved algorithm, two features of the original CSA are modified: (I) the awareness prob- ability (AP) and (II) the random perturbation. With the purpose to enhance the exploration–exploitation ratio, the fixed awareness probability (AP) value is replaced (I) by a dynamic awareness probability (DAP), which is adjusted accordingtothefitnessvalueofeachcandidatesolution.TheLévyflightmovement is also incorporated to enhance the search capacities of the original random per- turbation (II) of CSA. In order to evaluate its performance, the algorithm has been tested in a set offour optimization problems which involve induction motors and distribution networks. The results demonstrate the high performance of the pro- posed method when it is compared with other popular approaches. InChap.4,acomparativestudybetweenmetaheuristictechniquesusedforsolar cells parameter estimation is presented. The comparison evaluates the solar cell models of one diode, two diodes, and three diodes. In the analysis, the solar cell models are evaluated considering different operation conditions. Experimental results obtained during the comparison are also statistically validated. Chapter 5 considers a nonlinear system identification method based on the Hammerstein model. In the scheme, the system is modeled through the adaptation of an adaptive network-based fuzzy inference system (ANFIS) scheme, taking advantageofthesimilitudebetweenitandtheHammersteinmodel.Toidentifythe parameters of the modeled system, the approach uses a recent nature-inspired method called the gravitational search algorithm (GSA). Different to most of existent optimization algorithms, GSA delivers a better performance in complex multimodalproblems,avoidingcriticalflawssuchastheprematureconvergenceto suboptimal solutions. To show the effectiveness of the proposed scheme, its modeling accuracy has been compared with other popular evolutionary computing algorithms through numerical simulations on different complex models. In Chap. 6, a methodology to implement human-knowledge-based optimization strategies is presented. In the scheme, a Takagi-Sugeno Fuzzy inference system is used to reproduce a specific search strategy generated by a human expert. Therefore, the number of rules and its configuration only depends on the expert experience without considering any learning rule process. Under these conditions, each fuzzy rule represents an expert observation that models the conditions under which candidate solutions are modified in order to reach the optimal location. To exhibit the performance and robustness of the method, a comparison to other well-known optimization methods is conducted. The comparison considers several standard benchmark functions which are typically found in the scientific literature. The results suggest a high performance of the proposed methodology. Preface ix Chapter 7 presents a recent metaheuristic algorithm called Neighborhood-based Consensus for Continuous Optimization (NCCO). NCCO is based on typical processespresentinmulti-agentsystems,suchaslocalconsensusformulationsand reactive responses. These operations are conducted by using appropriate operators that are applied in each evolutionary stage. A traditional metaheuristic algorithm considers in its operation the application of every operator without examining its finalimpact inthesearchingprocess.Incontrasttoothermetaheuristictechniques, the proposed method uses additional operators to avoid the undesirable effects produced by the over-exploitation or suboptimal exploration of conventional operations. In order to illustrate the performance and accuracy of the proposed NCCO approach, it is compared to several well-known, state-of-the-art algorithms over a set of benchmark functions, and real-world design applications. The experimental results demonstrate that NCCO’s performance is superior to the test algorithms. Chapter 8 presents a metaheuristic algorithm in which knowledge extracted during its operation is employed to guide its search strategy. In the approach, a self-organizing map (SOM) is used as extracting knowledge technique to identify the promising areas through the reduction of the search space. Therefore, in each generation, the scheme uses a subset of the complete group of generated solutions seensofartotraintheSOM.Oncetrained,theneuralunitfromtheSOMlatticethat corresponds to the best solution is identified. Then, by using local information of thisneuralunitanentirepopulationofcandidatesolutionsisproduced.Withtheuse oftheextractedknowledge,thenewapproachimprovestheconvergencetodifficult high multimodal optima by using a reduced number of function evaluations. The performance of our approach is compared to several state-of-the-art optimization techniques considering a set of well-known functions and three real-world engi- neering problems. The results validate that the presented method reaches the best balance regarding accuracy and computational cost over its counterparts. Guadalajara, Mexico Erik Cuevas Jorge Gálvez Omar Avalos