INTERNATIONAL SOUND AND VIBRATION DIGEST Published by THE INTERNATIONAL INSTITUTE OF ACOUSTICS AND VIBRATION (IIAV) Volume 2, Number 2 Date: November 11, 1996 Editor-in-Chief: Malcolm J. Crocker, Auburn University, USA Assistant Editor: Dragana Obradovic, Auburn University, USA Editorial Board: Duan-shi Chen, Jiao Tong University, Shanghai, CHINA Frank Fahy, ISVR, Southampton University, UK Jean L. Guyader, INSA de Lyon, FRANCE Colin H. Hansen, University of Adelaide, AUSTRALIA Hanno Heller, DLR, Braunschweig, GERMANY Nikolay Ivanov, Baltic State University, St. Petersburg, RUSSIA Finn Jacobsen, Technical University of Denmark, DENMARK G. Krishnappa, Institute for Machinery Research, NRC, CANADA Conny Larsson, Uppsala University, SWEDEN Leonid M. Lyamshev, Andreev Acoustics Institute, Moscow, RUSSIA Dan B. Marghitu, Auburn University, USA Eric Marsh, Penn State University, USA M.L. Munjal, Indian Institute of Science, Bangalore, INDIA David E. Newland, The University of Cambridge, UK Michael P. Norton, University of Western Australia, AUSTRALIA A. Selamet, Penn State University, USA Andrew F. Seybert, University of Kentucky, Lexington, USA Jan W. Verheij, TNO, Delft, THE NETHERLANDS. Current number of subscribers: 1035 To send a submission to the IS&V DIGEST or to subscribe or unsubscribe, send information by E-mail to yanas@eng.auburn.edu. TODAY'S DIGEST CONTENTS ITEM 1. INTRODUCTION. ITEM 2. IN MEMORIAM: Prof. Dr. rer. nat. Manfred Heckl. ITEM 3. IIAV AFFILIATED SCIENTIFIC SOCIETY: The Brazilian Acoustical Society (SOBRAC). ITEM 4. BIBLIOGRAPHY: Active Noise and Vibration Control. ITEM 5. ESTABLISHMENT OF CENTER FOR APPLIED CONTROL (CAC). ITEM 6. ANNOUNCEMENT AND CALL FOR PAPERS: INTERNOISE 97. ITEM 7. ISVR CONFERENCE ANNOUNCEMENTS: Southampton, UK, 14 -17 and 21-25 July 1997. ITEM 8. ANNOUNCEMENT: Euromech Colloquium 369, 23-26 September 1997. ITEM 9. INTERNATIONAL SYMPOSIUM ANNOUNCEMENT: Bangalore, INDIA, 15-17 December, 1998. ITEM 10. INTERNATIONAL CONGRESS ANNOUNCEMENT: St. Petersburg, RUSSIA, September 1997 . ITEM 11. BOOK REVIEW: Multibody Dynamics with Unilateral Contacts. ITEM 12. BOOK REVIEW: Principles of Vibration and Sound ITEM 13. BOOK REVIEW: Engineering Vibration Analysis with Application to Control. ITEM 14. TECHNICAL PAPER: Genetic Algorithm Adaptation of Filter Weights for Active Noise and Vibration Control. ***************************************************************************** ITEM 1. INTRODUCTION. ***************************************************************************** We are pleased to bring you the next issue of the ISV DIGEST. The DIGEST is now being published by the INTERNATIONAL INSTITUTE OF ACOUSTICS AND VIBRATION (IIAV). The digest will be free of charge to members of the IIAV, but starting in 1997 there will be a moderate charge for non-members who wish to continue to receive the DIGEST. As explained in ITEM 2 of the last DIGEST, the IIAV is a new democratic international scientific society and membership as member or associate is open to individuals from all countries. So far over 160 individual members have joined from 33 different countries. In addition 14 scientific societies have become affiliated with the IIAV. On behalf of Sir James Lighthill, of London, England, the first President of IIAV and the officers and directors, it is my pleasure to invite you to consider becoming a member if you are not already. Members, as part of their annual dues of $80, will receive not only the DIGEST but also a refereed journal the INTERNATIONAL JOURNAL OF ACOUSTICS AND VIBRATION (IJAV) which will initially be published quarterly starting at the end of 1996. Non-IIAV members and libraries can subscribe to the JOURNAL (IJAV) for $100 per year by airmail, post paid. As the IIAV membership grows, additional services will be provided for IIAV members. It is planned to make full use of all modern electronic means of communication to assist members in the future. Complete details about IIAV are given in ITEM 2 of the last DIGEST. We were very saddened to receive the unexpected news of the death of Manfred Heckl in August 16, 1996. For those of us who were privileged to know him we will miss a scientist with a broad and deep technical knowledge who took a deep pleasure in his subject: acoustics. But our much greater loss is of a warm, wonderful, modest human being of great kindness. As discussed in the last DIGEST, plans are already well underway for the Fifth International Congress on Sound and Vibration. It will be held at the University of Adelaide, at Adelaide in South Australia, December, 15-18, 1997. The General Chairman is Professor Colin Hansen. He is planning an excellent technical program and we hope that you will put this date on your calendar and make plans to attend. We are sure that you will enjoy a genuine Australian welcome with great hospitality. If you are interested in attending the Fifth Congress, Colin Hansen will be pleased to receive your 250 word abstract by mail: Department of Mechanical Engineering University of Adelaide, GPO Box 498, Adelaide, 5005, Australia, FAX: 61-8-303-4367 or e-mail: chansen@mecheng.adelaide.edu.au. The proceedings of the previous four congresses are still available. Details were given in the last DIGEST. If you are interested in further information, please write to yanas@eng.auburn.edu. We are pleased to inform you that back issues of the ISV Digest are available on the Auburn University Gopher (Academic Areas and Departments-College of Engineering-Department of Mechanical Engineering). Beginning November 18, back issues of the ISV Digest will be available on Auburn University web under http://www.auburn.edu/isvd. We hope that you are finding the ISV DIGEST useful. We should be interested to hear your comments and to receive news items and technical information and articles for inclusion in the next issue. We hope to bring out the next issue of the DIGEST early in 1997, so don't forget to send your items soon! These should be sent to yanas@eng.auburn.edu. Malcolm J. Crocker Editor-in-Chief ***************************************************************************** ITEM 2. IN MEMORIAM: Prof. Dr. rer. nat. Manfred Heckl. ***************************************************************************** Professor Dr. rer. nat. Manfred Heckl has left us unexpectedly. Our memories of recent daily work with him are so vivid that the pain of his loss is more deeply felt. We disciples and colleagues remember first of all the human qualities of our teacher and "Institutsvater." He was a benevolent man with a big heart, known for his helpfulness and care for others. Though in the limelight of international science, he remai ned a man of the utmost modesty. It was his work that was most important to him, not his reputation. The many recent letters of condolence from his students and coworkers illuminate like flashbacks the character of this man as a true friend. Professor Heckl's scientific achievements enjoy world-wide esteem; his reputation was international. This is not only because of the large number of his publications, including important works like "Structure-Borne Sound," but also because his papers ref lect an unending search for the essential, the center where everything converges. In all his papers, theory was never an end in itself but was always a tool used to reach his primary goal, namely the clarification of physical and technical phenomena. Yet he valued theory and was fond of the expression, "There is nothing more practical than a sound theory." The presentation of his knowledge which he did with the facility of a virtuoso, made his lectures a memorable experience. Listeners were captivated not only by his experimental demonstrations and his well-known computerized movies, but equally by his imp ressive wealth of experience. The audience would follow his examination of the problem in question in silent suspense; they experienced the solution with pure joy. Apart from his professional qualities it was his humanity that earned him important and prestigious positions at his University, the Technical University of Berlin, where he was respected as an integrating force for the various groups. Professor Heckl se rved for many years as dean and prodean of "his" department of Environmental Technology; as a member of the Academic Senate he assumed, in addition, responsibilities benefiting the entire university. Out of all his endeavors and deeds one true Heckl trait becomes apparent: a deeply felt responsibility and an extraordinary sense of duty. These informed his insightful approach to physics and to practical problems as well as his friendly engagement with the people around him. A great man has left us, but not his example. M. Moeser Professor Technical University, Berlin (Edited, 10/24/96 RC.) ***************************************************************************** ITEM 3. IIAV AFFILIATED SCIENTIFIC SOCIETY ***************************************************************************** The Brazilian Acoustical Society (SOBRAC) The Sociedade Brasileira de Acustica (SOBRAC) was founded on November, 1984. Several technical meetings and other related functions had, however, occurred long before this date, including various activities at universities, research institutes, consultancies, manufacturers, noise and vibration product suppliers, and equipment manufacturers. SOBRAC now has over 600 members, including 450 individuals, 60 students, and 2 life members. SOBRAC is also supported by 85 industrial organizations. In accordance with the SOBRAC by-laws, SOBRAC board directors are elected to serve a two-year term. SOBRAC presidents have included Prof. Jules Slama from the Federal University of Rio de Janeiro (1984-1985), Prof. Joao Barring from Sao Paulo Institute of Research (1986-1987), Prof. Samir N. Y. Gerges from the Federal University of Santa Catarina (1988-1989), Mr. Mario Pimental from Vibranihil (Vibration Isolator Manufacturers) who was elected for two terms (1990-1993), and the current SOBRAC president Prof. Samir N. Y. Gerges (1994-1997) also elected for two terms. SOBRAC's main objective is to facilitate the sharing of information in acoustics and vibration between researchers, manufacturers, consultants, and users. This is achieved by holding regular annual meetings, symposia, and through the SOBRAC publications. SOBRAC publishes a biannual technical journal, Revista de Acustica e Vibracoes, which covers the major fields of noise and vibration. About 100 pages are published in each issue. Two thousand copies are distributed to society members, others in Brazil, and to individuals abroad. SOBRAC has played an active role in the formulation of the Brazilian "Law of Silence" passed in 1990 and the "Vehicle Noise Law" adopted in 1993. SOBRAC has representatives at the Brazilian Institute of Standards and in other societies related to safety at the workplace and acoustic comfort. SOBRAC has several active groups: 1) the Vehicle Noise Group which has held a symposium every other year since 1991 in Sao Paulo, 2) the Building Acoustics Group which holds a joint meeting with the Thermal & Ergonomic Groups of other societies and 3) a Hearing Conservation Group which works with safety medical officer groups of other societies such as those comprised of Audiologists, Safety Officers, and Industrial Medical Officers. SOBRAC has held annual meetings since 1980 in Sao Paulo, Rio de Janeiro, Santa Maria and Florianopolis. The most recent meeting was held during April, 1995, in Sao Paulo. This year the meeting is in Rio de Janeiro from 4-6 December 1996. It is planned to hold the first IberoAmerican Conference in Florianopolis in 1998, together with Argentina, Chile, Peru and Spain. SOBRAC has played a leading role in the formation of the Federacion IberoAmericana de Acustica (FIA) a federation of the acoustical societies of Argentina, Brazil, Chile, Mexico, Peru and Spain. Both SOBRAC and the FIA have become affiliated with the International Institute of Acoustics and Vibration (IIAV). ***************************************************************************** ITEM 4. BIBLIOGRAPHY ***************************************************************************** Bibliography on Active Noise and Vibration Control. The second supplement to the annotated reference bibliography "Active Noise and Vibration Control," compiled by D. Guicking at the University of Goettingen (Germany) has appeared in January 1996. The 291-page 3rd edition of February 1988, the 334-page 1st Supplement of August 1991, and the 447-page new supplement list a total of 5653 citations from 46 countries, covering the period from 1878 through 1994. The citations comprise Author(s), Title, Source, and Keywords (typically about 12), and are listed in alphabetical order of the first author. Access to specific references is facilitated by an Author Index and a Keyword Index. Text and explanations are in English. Copies are available from Dr. D. Guicking Drittes Physikalisches Institut Universitaet Goettingen Buergerstr. 42-44 D-37073 Goettingen (Germany) phone: +49-551-397727 or -397713 FAX: +49-551-397720 E-Mail: guicking@physik3.gwdg.de To cover printing and mailing expenses, a payment is requested the amount of which depends on the mailing fee and this can be determined by an inquiry to the above address. ***************************************************************************** ITEM 5. ESTABLISHMENT OF CENTER FOR APPLIED CONTROL (CAC). ***************************************************************************** Duke University in Durham, North Carolina is where the Center for Applied Control (CAC) has been recently established to: 1) Conduct Active Control research in a host of fields spanning acoustics, aeroacoustics, aeroelasticity, structural dynamics, nonlinear systems, pattern recognition, and seismic structural vibration. 2) Seek and develop industrial affiliates with pertinent applications. 3) Develop a core curriculum to prepare graduate students for research and academic careers in the above fields. The Duke University CAC also seeks to serve as a base of expertise for faculty members within the School of Engineering and external to the School to build interdisciplinary research ties. The center makes efficient use of faculty and laboratories to develop a program which otherwise would not be possible emanating from a single academic department. Duke University was recently awarded a new Sun UltraSpark 170E computer system to assist in launching the new center. Members of the group are currently in the process of setting up web pages as a resource of information. -- Research Projects -- Funded research programs and laboratories in the Departments of Civil and Environmental Engineering, Electrical and Computer Engineering, Computer Science, and Mechanical Engineering and Materials Science provide the core foundation for graduate programs supported by the Center. Also the Adaptive Systems and Structures Laboratory, the Acoustics Laboratory, the Machine Intelligence and Fuzzy Logic Laboratory, and the Seismic Response Control Laboratory are all affiliated with the CAC. Fundamental research activities in Applied Control currently underway are: Active Control of Aeroelastic Systems Airfoils in fans, compressors and turbines are subjected to a variety of unsteady structural dynamic and aerodynamic forces which produce undesirable airfoil vibration. These vibrations cause airfoils to fatigue -- reducing their useful life or even to fail catastrophically as in the case of the aeroelastic instability known as flutter. Fast and efficient computer models of the unsteady aerodynamic behavior of airfoils in turbomachinery are being developed for use in designing Active Control systems of aeroelastic phenomena. Particular interest is paid to the development of so-called reduced order models of unsteady aerodynamic flows. Using this approach the dominant eigenmodes of the flow about an airfoil, wing or cascade of airfoils are computed and used as basis functions to provide efficient reduced order models. Principal investigators are Profs. Kenneth Hall and Earl Dowell. Active Control of flexible aeroelastic surfaces provides opportunities for unexplored avenues and alternative solutions to existing technology problems. Control of instabilities associated with flutter will affect stability of aerodynamic surfaces as well as sound radiation. A continuing series of both experimental and theoretical studies over the last ten years have better defined the aeroelastic behavior of rotor blade systems and airfoils, with special emphasis on the effects of structural and aerodynamic nonlinear physical phenomena. Such phenomena are most prevalent at high performance and often high noise and high vibration conditions.. This research is being conducted by Prof. Earl Dowell and Robert Clark. Active Structural Vibration & Acoustic Control This project is devoted to the integration of actuators, sensors, and algorithms for the control of dynamic systems. Current research involves Active Control of structural vibrations due to the interaction between aerodynamic loads and the structure. The goal is to control the flutter in wings and thin panels used to construct aircraft. Active Structural Acoustic Control is being investigated for controlling interior noise within the fuselage of aircraft. New transduction devices for ASAC are currently being developed and applied to structures for feedback and adaptive feedforward control approaches. Additionally, current studies focus on the physics of structural acoustic coupling to interior cavities in order to determine methods of controlling broadband flow noise associated with turbulent boundary layers. These projects are directed by Prof. Robert Clark. Active Control of Seismic Structures This project is concerned with the suppression of earthquake-induced structural motions using advanced controllable materials, such as electrorheological (ER) and magnetorheological (MR) materials. Research is focused on: a) the analysis, design, fabrication, testing and modeling of devices which incorporate ER and MR materials. b) the synthesis of vibration control rules which make use of the properties of ER and MR devices. c) the testing of these control rules using small (1:5 to 1:3) scale models. The Seismic Response Control Laboratory (SRCL) has a high-bandwidth, high-force, controllable shaking table assembly, modern sensors, data acquisition and digital signal processing equipment. This effort is under the direction of Prof. Henri Gavin. Nonlinear Active Control New tools and techniques from nonlinear geometric control theory and exterior differential systems are needed to investigate the control of onholonomic and nonlinear systems. Specifically, current research includes: a) Developing obstacle-avoidance path planning algorithms and interactive steering simulations to test these algorithms. b) Stabilizing the system around the resulting trajectories and finding actual implementations that work well. c) Generating trajectories for systems with drift. d) Employing exterior differential system techniques to develop nonlinear normal forms for use in control theory. e) Investigating how to convert to these normal forms by using feedback transformations and how to control systems in these normal forms. Applications include Active Control of nonlinear locomotion systems such as wheeled robotic vehicles, aircraft, swimming robots, and snake- like robots. This project is directed by Prof. Linda Bushnell. Fuzzy Logic Active Control Applications Control systems are needed in almost every aspect of our current lives, and intelligent, adaptive, decision making control systems are required to operate under varying operational and environmental conditions. Fuzzy logic provides an overall decision making strategy amenable to Active Control system design, and the focus of research in this project is to convey the opportunities afforded with the integration of fuzzy logic, artificial intelligence, neural network and genetic algorithms (however, the main emphasis is on fuzzy logic) into modern control systems. The project is directed by Prof. Paul Wang. Active Magnetic Bearings A new technology, now being explored and gaining acceptance, is that of Active Magnetic Bearings, which replace conventional bearings and suspend the rotor by forces of magnetic attraction. Relatively simple active feedback can accomplish the goal of simple levitation, but more advanced methods are needed for vibration control of flexible shafts. Many issues must be explored, including those of observability and controllability in flexible rotor dynamics, options for sensor type and placement, relative merits of linear and nonlinear control, combinations of passive and active methods, integration of digital control, and adaptive feedforward and feedback methods. Prof. Josiah Knight is the project director. -- Project Management -- Each research program is managed by a CAC faculty member who serves as the Principal Investigator (PI) or Co-PI with faculty members in the School of Engineering. Opportunities for collaboration are widespread as control theory can be applied to many fields of study. Understanding the physics of each field of application is critical to practical control implementation and often motivates collaborative efforts. -- Education Goals -- Education is one of the most important aspects of the CAC as it serves to enhance the quality of research through a well developed and thorough graduate curriculum. The comprehensive program within the Duke University CAC ensures that graduate students obtain a fundamental education in applied and theoretical control of linear and nonlinear systems. -- Corporate Sponsors -- Another primary focus of the CAC is to establish and maintain strong ties with industrial affiliates interested in the forefront of research in Applied Control. The Lord Corporation, Scientific Applications International Corporation (SAIC) and the Cessna Aircraft Company have all written letters of support for the foundation of this new Center for Applied Control, and all are serving as industrial affiliates of the Center. Results of research programs are presented at an annual workshop on the Center activities, and industrial affiliates are openly given an opportunity to discuss and advise researchers on topics of particular need and interest to their respective application areas. For further information, contact Professor Robert Clark, Director of the Center for Applied Control, at 919-660-5435 or . ***************************************************************************** ITEM 6. ANNOUNCEMENT AND CALL FOR PAPERS: INTERNOISE 97. ***************************************************************************** INTER-NOISE 97, the 1997 International Congress on Noise Control Engineering, will be held at the Technical University of Budapest, Hungary from 1997 August 25 to 27. The Congress is sponsored by the International Institute of Noise Control Engineering, and is being organized by the Acoustical Commission of the Hungarian Academy of Sciences and the Hungarian Scientific Society for Optics, Acoustics, Motion Pictures and Theatre Technology. The theme of INTER-NOISE 97 is Help Quiet the World for a Higher Quality Life. Congress Secretariat: Scientific Society for Optics, Acoustics, Motion Pictures and Theatre Technology (OPAKFI) H-1027 Budapest, Foe u. 68. Hungary Phone/Fax: (Int) + 36 - 1 - 2020452, (Int) + 36 - 1 - 2018843 E-mail: in97.opa@mtesz.hu Internet: http://www.mmt.bme.hu/events/inter-noise97/c4papers.html CONTRIBUTIONS INVITED Proposals for papers in all areas of noise control engineering are welcome. Abstracts of papers proposed for presentation at INTER-NOISE 97 must be received by the General Secretary not later than 1996 November 30. The abstract should be approximately 250 words in length, and must be submitted in a special format. (For a copy of the format, please contact the Secretariat.) Technical papers in all areas of noise and vibration engineering will be considered for presentation at the congress. The following scientific and technical areas are of particular interest: - Active Noise and Vibration Control - Acoustical Simulation, Modeling, Identification and Prediction - Aircraft and Airport Noise Control - Building Acoustics, Sources and Transmission of Noise and Vibration - Construction Equipment Noise and Vibration Control - Physiological and Psychological Effects of Noise - Education, Public Relations for Noise and Vibration - Flow Acoustics, Fans, Ducts and Ventilation Systems - Low Frequency Noise and Vibration - Measurements and Analysis of Systems - The Practice of Noise and Vibration Control - Numerical Methods in Acoustics and Vibroacoustics - Non-linear Noise and Vibration Signal Processing - Outdoor Sound Propagation - Policy of Noise and Vibration Control - Standards, Regulations - Structural Transmission - Subjective Evaluation: Sound Quality, Sound Quality Engineering - Vehicle Noise Control - Computer Methods in Noise and Vibration - Outdoor Concert Sound - The Social Response to Environmental Noise and Vibration - Vibration Damping - Binaural Technologies - Evaluation of Impulse Noise EQUIPMENT EXHIBITION A major acoustical equipment, materials and instrument exhibition will be held in conjunction with INTER-NOISE 97. ----------------------------------------------------------------- ACTIVE 97, the l997 International Symposium on Active Control of Sound and Vibration, will be held at the Technical University of Budapest, Hungary from August 21 to 23, 1997. It precedes INTER-NOISE 97. The symposium is being organized by the Acoustical Commission of the Hungarian Academy of Sciences and the Hungarian Scientific Society for Optics, Acoustics, Motion Pictures and Theatre Technology in cooperation with the Institute of Noise Control Engineering (INCE/USA) and sponsored by the European Acoustics Association (EAA). In continuation of a series of meetings being held in Blacksburg, Virginia and Newport Beach, California from 1991 to 1995, ACTIVE 97 will be the first international conference on active control in Europe. ACTIVE 97 is planned to present six plenary keynote lectures covering current aspects and latest developments of active noise and vibration control. In addition, approximately l50 technical papers are expected to be presented. The conference chairmen is Fulop Augusztinovicz. Symposium Secretariat: Scientific Society for Optics, Acoustics, Motion Pictures and Theatre Technology (OPAKFI) H-1027 Budapest Foe u. 68., Hungary Phone/Fax: (Int) + 36 - l - 2020452, (Int) + 36 - l - 2018843 E-mail : a97.opa@mtesz.hu Internet: http://www.mmt.bme.hu/events/active97/c4papers.html Technical papers in all areas related to the active control of sound and vibration are welcome. Abstracts of papers proposed for presentation at the symposium must be received by the General Secretary not later than December 1, 1996. The abstract should have approximately 250 words in length, and must be submitted in a special format. (For a copy of the format, please contact the Secretariat.) AREAS AND SUBJECTS OF INTEREST Active noise control Active vibration control Active control of structurally radiated sound Application of active systems to industrial noise control Application of active systems to interior noise control Application of active systems to traffic noise control Application of active systems to marine noise control Application of active systems to vibration control Active vibration isolation Active control of building vibration Active control of fluid flow Active sound field reconstruction and modification Implementation and application of adaptive and smart structures Definition and application of psychoacoustic criteria to the active control of sound fields Feedforward control in active systems Feedback control in active systems Adaptive controllers in active systems Advanced controllers in active systems Controller hardware for active systems Actuators for active systems Sensors for active systems ---------------------------------------------------------------------- Fulop Augusztinovicz E-mail: fulop@hit.bme.hu Assoc. Prof., Technical Univ. of Budapest, Dept. of Telecommunications Sztoczek u. 2, H-1111 Budapest, Hungary Mail: H-1521 Budapest Tel: + 36 1 463 3246 Fax: + 36 1 463 3266 ***************************************************************************** ITEM 7. ISVR CONFERENCE ANNOUNCEMENTS ***************************************************************************** SIXTH INTERNATIONAL CONFERENCE ON RECENT ADVANCES IN STRUCTURAL DYNAMICS 14-17 July 1997 at the Institute of Sound and Vibration Research, University of Southampton, England. FOURTH INTERNATIONAL CONFERENCE ON NATURAL PHYSICAL PROCESSES RELATED TO SEA SURFACE SOUND 21-25 July 1997 at Chilworth Manor, Southampton, England. Further details of both conferences from Maureen Strickland at the ISVR, University of Southampton, Southampton SO17 1BJ, England. Tel: +44 1703 592294; fax: +44 1703 593190; e-mail: mzs@isvr.soton.ac.uk. ***************************************************************************** ITEM 8. ANNOUNCEMENT ***************************************************************************** EUROPEAN MECHANICS SOCIETY EUROMECH COLLOQUIUM 369 FLUID-STRUCTURE INTERACTION IN ACOUSTICS 23-26 September 1997 Scientific Committee D.G. Crighton, P.J.T. Filippi, D. Habault, D. Juve, A.H.P. Van der Burgh, J. Wauer Chairman: Adriaan H.P. van der Burgh Co-Chairmain: Paul J.T. Filippi FIRST ANNOUNCEMENT AND CALL FOR ABSTRACTS Dead-line for submitting abstracts - 31 January 1997 Scope of the colloquium The theme of the Colloquium concentrates on those areas where the interaction between sound and vibrating structures is significant. Topics covered are radiation and scattering from structural elements and fluid-loaded structures, including structures excited by flowing media. The interaction concerns two main classes of coupling: the strong and the weak coupling which correspond for instance with the coupling of a structure with a liquid respectively with a gas. Both theoretical as experimental research is included in the theme. In the real world applications are related for instance to vibrations, engine and flow noise inside vehicles, trains and aeroplanes as well as noise radiated by pipes excited by a turbulent flow. Acoustic transmission of information such as speech or musical sounds is an other field of application covered by the theme. Traditionally the Euromech Colloquia are informal, there will be no parallel sessions or formal published proceedings. The number of participants will be limited to approximately 60. Participation in the colloquium by invitation will have priority. Young researchers are encouraged to apply for participation. TOPICS Analytical concepts and methods - Exact and perturbation methods - Fundamental model problems Numerical models and methods - BEM and FEM, also combined - Statistical energy analysis - Evaluation of numerical results with analytical solutions and - approximations - Validation with experimental results Structure excitation by flows - Wall pressure fluctuations beneath unsteady or turbulent flows - Flow induced vibrations of curved and planar structures - Determination of the resulting sound radiation Experimental studies - Visualization techniques - Radiation and transmission of sound by vibrating structures including musical instruments Nonlinear fluid-structure interaction - Fluid-filled vessels and containers - Internal flow: piping, channels, turbomachinery and exchangers - External flow: offshore and marine problems, wind-induced phenomena Inverse problems and active control - Identification of elastic targets through acoustic diffraction as well as - sound emission mechanism - Passive and active control of radiated and transmitted noise If you wish to participate or would like further information, please contact: Adriaan H.P. Van der Burgh Delft University of Technology Faculty of Mathematics and Informatics P.O. Box 5031, 2600 GA Delft, The Netherlands Telephone: +31/15/2784420\ Fax: +31/15/2787209 e-mail: burgh@dv.twi.tudelft.nl or Paul J.T. Filippi Laboratoire de Mecanique et d'Acoustique 31 Chemin Joseph Aiguier 13402 Marseille CEDEX 20, France Telephone: +33/91/164068 Fax: +33/91/228248 e-mail: Filippi@lma.cnrs-mrs.fr ***************************************************************************** ITEM 9. INTERNATIONAL SYMPOSIUM ANNOUNCEMENT ***************************************************************************** PRELIMINARY ANNOUNCEMENT An International Symposium on " Designing for Quietness" will be on 15-17, December (Tuesday to Thursday) 1998, under the auspices of the International Institute of Acoustics and Vibration (IIAV) and the Acoustical Society of India. The Symposium will be held at the Center of Excellence for Technical Acoustics in the Department of Mechanical Engineering, Indian Institute of Science, Bangalore 560 012. Professor M. L. Munjal will be the General Chairman of the symposium. M. L. Munjal Professor and Head Department of Mechanical Engineering Indian Institute of Science Bangalore, India e-mal: munjal@iisc.ernet.in ***************************************************************************** ITEM 10. INTERNATIONAL CONGRESS ANNOUNCEMENT ***************************************************************************** FIRST INTERNATIONAL CONGRESS SCIENTIFIC PROBLEMS OF PRESERVATION OF HISTORIC CITIES ST.PETERSBURG, SEPTEMBER 1997 Congress announcement ********************* The Scientific Research Centre for Ecological Safety of the Russian Academy of Sciences and the State Museum of St.-Petersburg History will hold the FIRST INTERNATIONAL CONGRESS 'SCIENTIFIC PROBLEMS OF PRESERVATION OF HISTORIC CITIES' in St.Petersburg (September 1997). The congress is sponsored by: St.Petersburg Scientific Centre of Russian Academy of Sciences; Department of Culture, Department of Building construction and Architecture of St.Petersburg. General Chairman of Congress: director of Scientific Research Centre for Ecological Safety of Russian Academy of Sciences D.Sc. V.B. Donchenko. Scientific Co-chairmen: D.Sc.V.V.Sofronov and director of the State Museum of St.- Petersburg History N.L.Dementjeva. Executive director: S.V.Pestriakova. Congress objectives: -------------------- World famous cities are cultural centers with centuries-old history which exist as architectural and sculptural monuments and urban edifices. The centuries-old evolution of the cultural environment for the city population is reflected in many cultural layers and the necessity for their preservation is obvious. However, an urban environment is the domain of the life and activity for the city population and this environment undergoes natural and human influences. The increase in negative influence caused by industry and mechanized transportation, etc. leads to a process of intensive aging and accelerated decay in the city. Usually consequences of chemical pollution and extensive physical influences are considered from an ecological point of view, but city life and age also affect these influences to some degree. The Congress objective is the scientific analysis of possibilities for the preservation of historic cities. The Congress will encourage joint efforts of specialists in the humanities involved in the problem of preservation and development of urban culture with those of specialists in the natural sciences and engineering who are interested in solving the physical, mechanical and chemical problems of the urban environment. Congress topics *************** - historic problems of building construction; - city architectural archives and computer data bases; - stress analysis and destruction risk for old buildings; - geomechanics; - vibration and city noise; - heat fields; - city aerodynamics; - chemical influences; - long-term processes of building deformation; - damage tolerance; - additional problems ( please suggest). The Congress programme includes three general discussions on topics: "Humane problems of historic cities preservation", "Natural science problems of historic cities preservation" and "Ecological problems of historic cities" REQUEST FOR FURTHER INFORMATION Title.........Initials............Surname................ Organization............................................. Address.................................................. ........................................................ ......................................................... Telephone.....................Fax........................ Email..................................................... Please return completed form to: S.V.Pestriakova, CES RAN, Korpusnaya str. 18, 197042,St.Petersburg,Russia ----------------------------------------------------------- fax: 812 2354361 telephone: 812 2307836 e-mail: svetlana@srces.samson.spb.su ***************************************************************************** ITEM11. BOOK REVIEW ***************************************************************************** MULTIBODY DYNAMICS WITH UNILATERAL CONTACTS John Wiley & Sons, Inc., New York, 1996, 317pp., ISBN 0-471-15565-9. Price (U.S. $74.95). by F. Pfeiffer and C. Glocker As mechanical systems become more complex so do the mathematical models and simulations used to describe the interactions of their parts. One area of multibody theory that has received a great deal of attention in recent years is the dynamics of multiple contact situations that occur in continuous joints and couplings. Despite the rapid gains in our understanding of what occurs when continuous joints and couplings interact, until now there were no books devoted exclusively to this intriguing phenomenon. Focusing on the concerns of practicing engineers, Multibody Dynamics with Unilateral Contacts presents all theoretical and applied aspects of this subject relevant to a practical understanding of multiple unilateral contact situations in multibody mechanical systems. In Part 1, the authors provide an exhaustive review of the laws and principles governing the dynamics of unilateral contacts in multibody mechanical and technical systems. Among the topics covered are multibody and contact kinematics, the dynamics of rigid body systems, multiple contact configurations, detachment and stick-slip transitions, frictionless impacts, impacts with friction, and the Corner law of contact dynamics. In Part 2, the authors present numerous applications of the theories presented in Part 1. Each chapter in this part is devoted to a different law, theory, or model, such as discontinuous force laws, classical impact theory, Coulomb's friction law, and mechanical and mathematical models of impacts and friction. In addition, each chapter features several practical examples that allow engineers to observe the concepts described in action. Examples are drawn from a broad array of fields and range from hammering in gears as occur in a synchronous generator to impacts and friction as observed in a child's woodpecker toy, from a demonstration of classical impact theory using an automobile gear box example, to Coulomb's friction law as applied to a turbine blade damper. Multibody Dynamics with Unilateral Contacts is an indispensable resource for mechanical engineers working on all types of multibody systems and the friction and vibration problems that can occur in them. It is also a valuable reference for researchers studying nonlinear dynamics. Dan B. Marghitu Assistant Professor of Mechanical Engineering Auburn University e-mail: marghitu@eng.auburn.edu ***************************************************************************** ITEM 12. BOOK REVIEW ***************************************************************************** PRINCIPLES OF VIBRATION AND SOUND T.D. Rossing and N.H. Fletcher Springer-Verlag New York, Inc., 1995, 247 pp., ISBN 0-387-94336-6, Price (U.S. $34.50). This book is an outgrowth from the earlier book, Physics of Musical Instruments, by the same authors, published by Springer-Verlag in 1991. The present book presents the material of the first third of the earlier book, developed as a textbook suitable for courses in vibrations and acoustics. Thus, the book provides interesting and contains somewhat different material from that found in similar vibrations and acoustics books. The emphasis in this book is obviously on the basic physics of musical instruments, rather than on engineering applications such as underwater acoustics or noise and vibration as might be found in similar textbooks. The book begins with standard chapters on free and forced vibrations of simple systems, vibrations of continuous systems in one dimension, such as strings and bars and vibrations of two-dimensional systems: membranes and plates. The book continues with a chapter on coupled vibrating systems and another on non-linear systems. This concludes the first part of the book which deals with vibrating systems. The second part contains four chapters concerned with sound waves in air. The first chapter of part two contains discussions on standard topics such as plane and spherical waves, sound pressure level, intensity, reflection, transmission, absorption, and normal modes in cavities. The next chapter discusses sound radiation from spherical sources including monopole, dipole, quadrupole sources, line sources, and piston sources, etc. The last two chapters contain discussions on propagation in pipes and horns including conical and exponential horns, Bessel and compound horns. The book concludes with questions for each chapter and answers to odd numbered questions. The book is written at a level suitable for undergraduate students and others interested in the vibration and acoustic properties of musical instruments. The authors write in an authoritative manner and are obviously experts in their fields. They have presented the material in an accessible way for students with plenty of descriptive text as well as suitable equations. The practical examples given are normally those of stringed instruments or cavities in wind and other instruments, etc. This book can be recommended not only to all those interested in the acoustics of musical instruments. It will also be of interest to many other readers who have a general interest in vibration and acoustics theory. Malcolm J. Crocker Department of Mechanical Engineering Auburn University e-mail: mcrocker@eng.auburn.edu ***************************************************************************** ITEM 13. BOOK REVIEW ***************************************************************************** ENGINEERING VIBRATION ANALYSIS WITH APPLICATION TO CONTROL SYSTEMS C.F. Beards Halsted Press, 1996, 426 pp., ISBN 0-470-23559-4. Price (U.S. $44.95). This book is designed not only for the use of students but practicing engineers and scientists as well. It contains six chapters which include 75 worked examples. In addition it contains 130 problems together with detailed worked solutions to some of these problems. One interesting and somewhat unusual feature of the book is that it contains extensive discussions on vibration theory and also on automatic control systems. After the brief introductory chapter, chapter two deals with one degree of freedom systems. The discussion includes not only translational vibration, but torsional vibration as well. Both free and forced vibration theory is worked out. The cases of viscous, Coulomb and hysteretic damping are all considered. Vibration isolator design is also described. Chapter three deals with the vibration of systems with more than one degree of freedom. Both free and forced vibration and the ideas of receptances, impedance, and mobility are described. This chapter also includes discussions of the matrix method, orthogonality of the principle modes and Lagrange's equation. Chapter four deals with vibration of continuous systems. The chapter contains discussions on wave motion including transverse, longitudinal, and torsional vibrations. Extensive discussion is included on transverse vibration such as the vibration of uniform beams, with and without the presence of discrete bodies and the whirling of shafts. The analysis of continuous systems using Rayleigh's energy method also receives coverage. Chapter five is devoted to automatic control systems, including simple hydraulic servo systems, modifications to simple hydraulic servo systems, the electric position servomechanism and control system frequency response, among other related topics. The book concludes with chapter six devoted to thirty pages of problems and chapter seven provides detailed answers to selected problems. Also included are a bibliography and index. This is an attractive book because it contains numerous practical worked examples and extensive worked solutions to the problems. The book is written by a knowledgeable author, well established in his field who has wide practical experience. The text is presented at a level that will be useful not only to students, but practicing engineers and scientists as well. Malcolm J. Crocker Department of Mechanical Engineering Auburn University e-mail: mcrocker@eng.auburn.edu ***************************************************************************** ITEM 14. TECHNICAL PAPER ***************************************************************************** GENETIC ALGORITHM ADAPTATION OF FILTER WEIGHTS FOR ACTIVE NOISE AND VIBRATION CONTROL Colin H. Hansen Department of Mechanical Engineering University of Adelaide South Australia 5005. e-mail: chansen@mecheng.adelaide.edu.au In this article, the results of some new research currently being undertaken on the adaptation of controller filter weights using a genetic algorithm are given. Although the convergence rate of this algorithm is much slower than that of the standard LMS algorithm, the advantages of the genetic algorithm are that it can be used to adapt non-linear filters, no estimate of the cancellation path transfer function is needed to make it work (thus the error signal need not be correlated with the reference signal) and it is inherently very stable. The algorithm to be discussed here is virtually independent of the type of filter structure which is used, and the best type of filter structure for a particular problem can be selected easily by using trial and error. The principal disadvantage of the genetic algorithm is that it is relatively slow, limiting its usefulness to relatively steady systems or systems which vary slowly with time. This is a result of the averaging time required for performance measurement which is at least half the period of the lowest frequency signal encountered in the error signals, which is the lowest frequency to be controlled if a suitable high pass filter is used. The genetic algorithm is an optimization/search technique based on evolution, and is essentially a guided random search. Use of the genetic algorithm enables any filter structure to be treated as a "black box" which processes reference signals to produce control signals, based on different sets of filter weights. Basic genetic algorithm operation requires the testing of solutions (sets of filter weights), which involves loading the filter weights into the filter and subsequently evaluating the performance of the filter in minimizing a cost function based on the error sensor outputs. The genetic algorithm in essence combines high performance solutions while also including a random search component. ALGORITHM IMPLEMENTATION Implementation of the genetic algorithm described here has three basic stages: fitness evaluation, selection and breeding. Fitness evaluation requires the testing of the performance of all individuals in the population. Here an individual is considered to be a separate set of filter weights, with the fitness of the individual being a measure of the filter's performance when these weights are being used for filter output calculation. The population then consists of a collection of these individuals. Selection involves killing a given proportion of the population based on probabilistic 'survival-of-the-fittest'. Killed individuals are replaced by children, which are created by breeding the remaining individuals in the population. Typically 70% of the population are killed, with the remaining 30% forming the mating pool for breeding. For each child produced, breeding first requires probabilistic selection of two (possibly the same) parent individuals, with fitter individuals being more likely to be chosen. The probability of selection is high for parents of "good" fitness and low for parents of "poor" fitness. For optimal results, it is best to vary the probability distribution depending upon the stage of convergence which the algorithm has reached. In a typical probability distribution there is a heavy selection bias (or high selection pressure) towards the fitter individuals in the beginning and less selection pressure towards the end of the convergence process. Application of the crossover and mutation operators on the parent pair produces the new child. The crossover operator combines the information contained in two parent strings (or two sets of filter weights) by probabilistic copying of information from either parent to each corresponding string element (or single filter weight) of the child being produced. In this case, the probability of copying a particular weight value to the child from either of the two parents is the same. Mutation introduces random copying 'errors' during the information copying stage of crossover, and gives the algorithm a random search capability. Mutation plays a minor role in the implementation of the genetic algorithm in standard optimization problems, in that it is used to replace lost bits in the binary encoding of the problem. As binary data have only two states, small mutation probabilities work well with the 'standard' implementation where data loss is minor. This is not the case in the implementation most suited to active noise and vibration control, where a weight string is used instead of binary encoding [7]. Here, mutation is necessary to maintain population diversity (differences between individuals) and also to allow 'homing in' on optimal solutions, as the population data corresponding to one weight in the string will not fully represent the weight's entire data range. However in practice it is necessary to place bounds on the allowed range of mutation (mutation amplitude), the optimal bounds being somewhat problem dependent. Two selection processes are carried out during the operation of the genetic algorithm, namely the choice of individuals to be killed, and the choice of parents during breeding. Both selection processes have been implemented using a simulated roulette wheel, where each segment (or slot) on the roulette wheel is allocated a size proportional to the individual's probability of being chosen (selection probability). Each spin of the roulette wheel results in one "winner" being selected. Selection probabilities are assigned such that low performance individuals are more likely to be killed, and such that high performance individuals are more likely to be chosen as parents for breeding. Selection without replacement is used for killing, where once an individual is chosen it is removed from the roulette wheel. For breeding, selection with replacement (no removal) is used for choosing the parents, hence the entire mating pool is used in the selection of each parent individual. Many aspects of the genetic algorithm used in standard optimization implementations have been changed to give the desired on-line optimization performance required for active noise and vibration control [7], as discussed in the subsections to follow. KILLING SELECTION INSTEAD OF SURVIVOR SELECTION Choosing individuals to be killed rather than those to survive allows higher survival probabilities to be realized for the higher performing individuals. This enables greater selective pressure (bias towards survival and breeding of the higher performance individuals) to be applied, which can be used to give faster convergence when high levels of mutation are used to sustain population diversity. Use of killing selection also allows the best performing individual to be assigned a killing probability of zero to ensure its survival. WEIGHT STRING INSTEAD OF BINARY ENCODING The "genetic code" of each individual is normally encoded as a binary string from the problem variables; in this case, it would imply that each weight would be coded as a binary string and the strings connected together to form a complete individual or set of weights, with the crossover and mutation operators working at the single bit level. Mutation of the upper bits of weight variables would result in large jumps in weight values when filter weights are encoded in this way, which significantly degrades on-line performance. To alleviate this problem in active noise and vibration control systems, a weight string is used, with the crossover and mutation operators applied using whole weight values as the smallest operational element. MUTATION PROBABILITY AND AMPLITUDE Application of mutation to whole weight values enables a limit to be placed on the deviation of filter weight values about their current values, which gives control over the spread of the filter's performance. Mutation is applied to all child string variables at a given probability (mutation probability, typically 20 to 30%). The weights chosen to be mutated are modified by a random change in value, which is limited to a specified range (mutation amplitude). For best results the mutation amplitude should be relatively high at the start of convergence and low towards the end. Typical values range from 15% of the maximum possible weight value (at the start) to 0.01% of the maximum possible weight value (at the end of convergence). RANK-BASED SELECTION (KILLING AND BREEDING) Rank-based selection removes the scaling problems associated with fitness proportionate selection (assigning selection probabilities proportional to fitness values), and gives exceptional control over selective pressure [18,19]. Rank- based selection, used by Whitley and Hanson [19] for breeding (parent selection) has been extended in the active noise and vibration control application to include killing selection. Selection probabilities, for both killing and breeding, are assigned based on the rank position of each individual's performance. This essentially means that the individuals are sorted into order from best to worst performance, then each allocated a fixed selection probability (probability of being chosen) based on their position in this list. The performance evaluation method used thus becomes irrelevant as long as the rank positions are the same (or similar). Separate (adjustable) probability distributions are used for killing and parent selection, with killing being more probable for lower ranked individuals and selection to be a parent being more probable for higher ranked individuals. UNIFORM CROSSOVER Uniform crossover nearly always combines the information of two parent strings more effectively than one or two point crossover [20]. One point crossover is where a position along the string is selected at random, and information is copied (to the child being created) from one parent for the first part of the child string and from the other parent for the second part. Similarly two point crossover involves selecting two points along the string, and copying from one parent between these two points, and from the other parent for the rest of the child string. In uniform crossover each position along the child string is produced by randomly copying from either parent, with both parents being equally likely to be chosen as the information source. For active noise and vibration control problems it has been found that it is best to use a modified form of uniform crossover [7], for which the probability of copying information from the lower ranked parent is supplied, and whole weight values are the smallest elements that are copied (compared to single bits for binary encoded strings). GENETIC ALGORITHM PARAMETER ADJUSTMENT As suggested by De Jong [21], adjustment of the operating parameters (probabilities, population size, etc) can improve the performance of the algorithm. The adjustable parameters used in the active noise and vibration control implementation discussed here (population size, survival ratio, killing and breeding rank-probability distributions, crossover probability, mutation probability, and mutation amplitude) provide good control over the stages of adaptation needed when good on- line performance is required. PERFORMANCE MEASUREMENT To evaluate the fitness of an individual (set of filter weights) in minimizing the error signal it is necessary to average the mean square error signal over a period of time greater than the period of the lowest frequency signal present. There should also be a delay of twice this between each fitness (or performance) evaluation to allow any transient effects resulting from implementation of the previous individual (set of weights) to subside. It is interesting to note that the genetic algorithm can handle any form of performance measure, including a measure of power or intensity, whereas previously discussed algorithms, because of the instantaneous nature of their cost functions, cannot use power or intensity error criteria easily. In multi-channel systems, the performance measure for the genetic algorithm would be the sum of the average square error measured by each error sensor. For applications involving "more important" and "less important" error sensors, it is easy to weight the signal from individual error sensors accordingly. IMPLEMENTATION EXAMPLE A single channel example, involving a beam fixed at both ends, will now be used to demonstrate the effectiveness of the genetic algorithm with three different types of filter structure. The first filter used was a linear FIR filter which was only capable of producing frequencies present in the reference signal, in this case a 133 Hz sinusoid. The second filter structure used was a non-linear polynomial filter which consisted of two 50 tap FIR filters, the inputs of which were the reference signal raised to the fourth and fifth powers respectively, with the control signal obtained by adding the filter outputs. Raising a signal to the fourth power creates a signal consisting of the 2nd and 4th order harmonics of the initial signal content. Similarly, raising a signal to the fifth power gives a signal with 1st (fundamental), 3rd, and 5th order harmonic content. Hence this polynomial filter, referred to as a P4P5 filter, can only produce harmonics (including the fundamental) of the reference signal up to the fifth order. The neural network based filter structure used had 50 taps, one hidden layer with 20 nodes, and one (linear) output layer node (designated 50x20x1). Four different transfer functions were utilized simultaneously in the hidden layer, with equal numbers (that is, five) of each type being used. The non-linearity in the control excitation was introduced by not attaching the control shaker properly to the beam. There was also considerable harmonic distortion in the primary excitation. Final converged vibration levels obtained using each of the three types of filter structure are shown in figures 6 to 8. The genetic algorithm adapted FIR filter gave a maximum of 12dB mean square error (MSE) reduction within 40 seconds. The P4P5 filter gave 12dB at 50 seconds, and a maximum of 36dB within 3 minutes. The 50x20x1 neural network filter gave 24dB at 50sec, 30dB in 6 minutes, and a maximum of 32dB MSE reduction within 15 minutes. Power spectra showing final converged vibration levels (measured at the error sensor) obtained using each of the three types of filter structure are shown in figures 6 to 8 respectively. The higher order harmonics present with no control are due to non-linear output from the primary electrodynamic shaker. A summary of the attenuation achieved at the harmonic peaks is given in Table 1. 1st 2nd 3rd 4th 5th 6th 7th 8th FIR 18 -32 -8 -5 4 -4 -3 316 P4P5 40 8 7 13 10 15 26 3729 NN 47 -2 12 -7 17 14 27 3326 Table 1. Attenuations achieved for each harmonic For the FIR filter case, attenuation of the fundamental peak at 133Hz is limited due to the introduction of the higher order harmonics by the non-linear control source. The P4P5 filter achieved the best overall reduction, with all harmonic peaks being attenuated. In comparison, the neural network filter structure has given greater control of the 1st, 3rd, and 5th order components, but has caused the 2nd and 4th order components to increase. Note that the presence of small quantities of higher order harmonics in the reference signal seen by the controller (due to harmonic distortion in the signal generator) have allowed the attenuation of higher order harmonics that would not normally be possible for the FIR (8th and 9th harmonics) and P4P5 (6th to 9th harmonics) filter structures when given a purely sinusoidal reference. REFERENCES 1. Wangler, C. and Hansen, C.H. (1994). Genetic algorithm adaptation of non-linear filter structures for active sound and vibration control. Proceedings of IEEE, ICASSP, Conference on Acoustics Speech and Signal Processing, April 18-22, pp. III505-III508. 2. Whitley, D. (1989). The genitor algorithm and selection pressure: Why rank based allocation of reproductive trials is best. In Proceedings of the 3rd International Conference on Genetic Algorithms. pp. 116-121. 3. Whitley, D. and Hanson, T. (1989). Optimizing neural networks using faster, more accurate genetic search. In Proceedings of the 3rd International Conference on Genetic Algorithms. pp. 391-396. 4. Syswerda, G. (1989). Uniform crossover in genetic algorithms. In Proceedings of the 3rd International Conference on Genetic Algorithms. pp. 2-9. 5. De Jong, K. (1985). Genetic algorithms: A 10 year perspective. In Proceedings of the 1st International Conference on Genetic Algorithms and Their Applications. pp. 169-177.