................................ A PPPPP NN N A A P P N N N A A P P N N N AAAAAAA PPPPP N N N A A P N NN A A P N N ...................................... THE APPLIED PROBABILITY NEWSLETTER The Official Communication of the Applied Probability Society of INFORMS Volume 19 Number 1 September 3, 1998 -------------------------------------------------------------------- CONTENTS OF THIS ISSUE: -------------------------------------------------------------------- COUNCIL NEWS MESSAGE FROM THE CHAIR APPLIED PROBABILITY SECTION BECOMES SOCIETY CALL FOR NOMINATIONS: NEW COUNCIL MEMBERS APPLIED PROBABILITY AWARDS 1998 ERLANG PRIZE AWARDED TO CHAO AND DAI CALL FOR NOMINATIONS: BEST PUBLICATION AWARD INFORMS CONFERENCES FIRST ANNONCEMENT: 10TH INFORMS APPLIED PROBABILITY CONFERENCE MINUTES OF MONTREAL BUSINESS MEETING TALKS FOR SEATTLE SPONSORED SESSIONS LIST OF SPONSORED SESSIONS FOR CINCINNATI CALL FOR SESSION CHAIRS FOR PHILADELPHIA OTHER CONFERENCES CALL FOR PAPERS: THE 38TH IEEE CONFERENCE ON DECISION AND CONTROL CONFERENCE LIST FOR FALL 1998 APPLIED PROBABILITY IN EUROPE INTRODUCING EURANDOM REPORT: EUROPEAN ROUND TABLE ON MAINTENANCE MANAGEMENT (PART 2) MISCELLANEOUS ANNOUNCEMENTS NEW JOURNAL: METHODOLOGY AND COMPUTING IN APPLIED PROBABILITY NEW BOOK ANNONCEMENT SOFTWARE ANNONCEMENT: QUEUEING NETWORK ANALYSIS TOOL OPENING: POST-DOCTORAL POSITION OPENING: CHAIR IN MANAGEMENT SCIENCE/OPERATIONS RESEARCH EDITOR'S COORDINATES -------------------------------------------------------------------- MESSAGE FROM THE CHAIR -------------------------------------------------------------------- Dear friends and colleagues, It is my very great pleasure to write to you as the last chair of the Applied Probability Section, and the first chair of the new Applied Probability Society of INFORMS. That's right, as of last month, our group is officially a society. The upgrade to society status gives us a boost in prestige and will help us broaden our membership base, especially among non-INFORMS members, practitioners, and researchers and practitioners outside of the US. Many thanks to our past chair, Larry Wein, who got the ball rolling on the upgrade, and whose leadership during the past two years provided much of the basis for our application's success. Congratulations to Xiuli Chao and Jim Dai for being awarded the Erlang prize! Xiuli and Jim are giving a tutorial on their work at the Cincinnati meeting next spring. Thanks to Peter Glynn, Marty Puterman, and George Shanthikumar for their work on the awards committee. Also thanks to Ward Whitt for agreeing to serve on the committee for the next three years. Mark your calendars for the next Applied Probability Meeting. It will be held next summer in Ulm, Germany. See the enclosed announcement. Speaking of meetings, it's not too early to start thinking about the next one, to be held somewhere in the US in 2001. Any volunteers? Seems like the New York City area would be an obvious location... At our next business meeting in Seattle, October 26, we will be electing two new council members (see the enclosed announcement). Being a council member is fun and not a huge amount of work, so please send in your nominations, especially self-nominations, to me before the Seattle meeting. I look forward to seeing many of you in Seattle! Rhonda Righter Department of Operations and Management Information Systems Santa Clara University Santa Clara, CA 95053 RRighter@SCU.edu (408) 554-4977 -------------------------------------------------------------------- APPLIED PROBABILITY SECTION BECOMES SOCIETY -------------------------------------------------------------------- The INFORMS board of directors approved the Application for Society Status of the Secion on Applied Probability on August 8, 1988. Our group's official title is now the Applied Probability Society of INFORMS. The decision to apply for society status was taken in April at the AP Section's business meeting in Montreal, then the application was promptly prepared by our new society chair, Rhonda Righter. Good work! The new by-laws for the society can be found on the web site of the Applied Probability Society, at . -------------------------------------------------------------------- CALL FOR NOMINATIONS: NEW COUNCIL MEMBERS -------------------------------------------------------------------- The Applied Probability Society of INFORMS seeks nominations for two new council members, for election at the next business meeting in Seattle, on October 26, 1998. The term of office is for two years. According to the Society's bylaws: Nominations for the Council will normally reflect the spectrum of disciplines and/or roles represented by Society members. The Council will constitute the governing body of The Society and will guide the officers in the administration of Society business. Please send your nominations, including self-nominations, to Rhonda Righter, email , before the Seattle meeting. -------------------------------------------------------------------- 1998 ERLANG PRIZE AWARDED TO CHAO AND DAI -------------------------------------------------------------------- (Editor's note: this announcement appeared in OR/MS Today, June 1998.) The Erlang Prize, presented by the Applied Probability Section of INFORMS, was awarded to Xiuli Chao for his wide range of contributions to applied probability and to Jim Dai for his significant research in applied probability. Most significant is an extensive study by Chao and his collaborators of the class of Markovian networks which exhibit procut-form stationary distributions. This class of networks encompasses Jackson and Kelly networks, networks and negative customers, and signaling networks. In other investigations, Chao used stochastic convexity to establish an important monotonicity property for doubly stochastic Poisson processes and has studied interchangeability and reversibility properties of tandem queues with finite buffers and blocking. In addition to an outstanding record of research and publications, Chao has served on the editorial boards of many leading journals and has co-authored several books on queueing networks and operations scheduling. As part of his dissertation research, Jim Dai developed efficient numerical methods for steady-state analysis of multi-dimensional reflected Brownian motions, which arise as heavy traffic approximations of queueing networks. Additionally, he established a close mathematical relationship between fluid network stability and queueing network stability. This connection has since become a standard tool in studying multi-class stability problems. Dai's recent work has focused on characterizing the global stability region for multi-class networks, and on the formulation of easily implemented scheduling rules that promote stability. -------------------------------------------------------------------- CALL FOR NOMINATIONS: BEST PUBLICATION AWARD -------------------------------------------------------------------- 1999 Best Publication Award Call for Nominations The Applied Probability Section of INFORMS invites nominations for its 1999 Best Publication Award. The Award, which recognizes outstanding contributions to Applied Probability, consists of a plaque and $1000, to be presented at the INFORMS meeting in Cincinnati, May 1999. To be eligible for the Award, the contribution (a book, paper, or set of papers) must meet the following criteria: 1. the contribution must be in Applied Probability; 2. it must have been published (in the case of a set, at least one member of the set) in the open literature during the interval 1995-1998; 3. it must be written in English, or have an accurate English translation available. The contribution will be judged according to: 1. the extent to which it advances the state of the art of Applied Probability; 2. the originality of the theory or methodology, or the degree to which it simplifies or unifies the existing theory or methodology; 3. the new areas of application that it opens; and 4. the clarity and excellence of the exposition. Nominations may be made by anyone, including the nominee. A nomination package consists of: 1. a detailed description of the contribution and its significance, and 2. a copy of the nominated publication(s). The nominator should send a separate nomination package by 31 January, 1999 to each of the Award committee members: Martin L. Puterman, chair Faculty of Commerce and Business Administration University of British Columbia 2053 Main Mall Vancouver, BC, V6T 1Z2 Canada Phone: 604-822-8388 Fax: 604-822-9574 Email: Peter W. Glynn Department of Operations Research Stanford University Stanford, CA 93405-4022 Email: Ward Whitt AT&T Research 180 Park Avenue Florham Park, NJ 07932-0971 Email: -------------------------------------------------------------------- FIRST ANNONCEMENT: 10TH INFORMS APPLIED PROBABILITY CONFERENCE -------------------------------------------------------------------- FIRST ANNOUNCEMENT 10th INFORMS Applied Probability Conference Ulm, Germany July 26 - 28, 1999 The general theme of the conference is ''Applied Probability in Computer and Communication Sciences, Economics and Engineering''. The objectives are twofold: to bring together researchers working in this field throughout the world and to take advantage of this meeting to enhance collaboration between applied probabilists within Europe at large. This meeting will be held on July 26 - 28, 1999, at the University of Ulm, Germany. The following people have accepted invitations to present plenary lectures: F. Baccelli (INRIA Sophia Antipolis, France) Stochastic Networks R.D. Gill (Utrecht University, The Netherlands) Survival Analysis and Spatial Statistics A.N. Shiryayev (Steklov Mathematical Institute Moscow, Russia) Finance and Stochastics The deadline for submission of invited sessions is February 1, 1999 and for submission of all abstracts, March 1, 1999. Further information about this conference is available on the web at www.mathematik.uni-ulm.de/ap99 It includes a list of conference topics, the program committee and instructions about paper submissions. If you wish to receive future announcements and reminders about deadlines, please send an e-mail to AP99@mathematik.uni-ulm.de Organizing Committee: Nicole Baeuerle, Andreas Frey, Uwe Jensen Ulrich Rieder, Volker Schmidt, Ulrich Stadtmueller University of Ulm Faculty of Mathematics and Economics D-89069 Ulm, Germany -------------------------------------------------------------------- MINUTES OF MONTREAL BUSINESS MEETING -------------------------------------------------------------------- Minutes of the Applied Probability Section's Business Meeting Montreal, 27 April 1998 Submitted by: Robert D. Foley We had a huge turnout at the meeting of the Applied Probability Section chaired by Larry Wein at the CORS/INFORMS Spring conference in Montreal. Bernard Lamond was commended for his excellent job in organizing a large number of interesting applied probability sessions. (Editor's note: a Queueing cluster was also organized by David Stanford.) The first topic of discussion was whether our group should consider becoming a society. John Birge, Vice-President of INFORMS subdivisions, described the procedure and pros and cons of becoming a society. After some discussion, a straw vote was held, and the results strongly favored continuing the process of becoming a society. Larry pointed out that Rhonda Righter, our incoming president, will be handling this issue. Next Marty Puterman had the honor of awarding the 1998 Erlang Prize. The committee decided to award the prize to two people. The recipients were Xiuli Chao of NJIT and Jim Dai of Georgia Tech. The next award our group will be offering is the best publication in applied probability. The call for nominations (above) is on our excellent website which is maintained by Bernard Lamond (Bernard.Lamond@fsa.ulaval.ca). Daniel Heyman then presented the Marcel Neuts Prize, which is awarded annually by the Applied Probability Foundation of America to the author of the Best Paper in the previous year's volume of Stochastic Models. This year's winners were Gagan Choudury, Avi Mandelbaum, Martin Reiman and Ward Whitt for their paper "Fluid and diffusion limits for queues in slowly changing environments". (Editor's note: this award is not under the responsibility of the Applied Probability Section, but its business meeting served as a convenient venue for it.) Ulrich Rieder announced the 10th INFORMS Applied Probability Conference which will be held at the University of Ulm in Germany from July 26th to 28th, 1999. The deadline for submission of invited sessions is Feb. 1, 1999, and the deadline for submission of abstracts is March 1, 1999. Acceptance notifications will be sent by April 1. An invited session proposal should include a letter from the session organizer containing the session title and a list of speakers and their abstracts. Each session will consist of four 30 minute talks. Session proposals and abstracts should be e-mailed to AP99@mathematik.uni-ulm.de. Please see the website for format information. The Fall 1998 INFORMS meeting will be in Seattle from Oct. 25th--28th, and Mark van Oyen (vanoyen@nwu.edu) is the applied probability cluster chair. The Spring 1999 INFORMS meeting will be in Cincinnati, and Tava Lennon Olsen (tlennon@engin.umich.edu) will be the applied probability cluster chair. The Fall 1999 INFORMS meeting will be in Philadelphia, and, as of the Montreal meeting, we are still in search of a cluster chair for applied probability. (Editor's note: Natarajan Gautam, email , will organize the Applied Probability Cluster for Philadelphia. See the CALL FOR SESSION CHAIRS below.) -------------------------------------------------------------------- TALKS FOR SEATTLE SPONSORED SESSIONS -------------------------------------------------------------------- (Editor's note: see also the abstracts on the conference web site at URL .) Conference: Seattle INFORMS, October 25-28, 1998 Applied Probability Cluster Organizer: Mark P. Van Oyen Applied Probability Sessions: 1. Algorithms for dynamic jobshops Chair: Dimitris Bertsimas Authors: Florin Avram, Dimitris Bertsimas, Jay Sethuraman; Michael Veatch; Dimitris Bertsimas and David Gamarnik 2. Yield Management, Chair: Dimitris Bertsimas Authors: Itir Karesman, Garrett van Ryzin; Ioannis Ch. Paschalidis and John Tsitsiklis; Dimitris Bertsimas and Gina Mourtzinou; 3. Stochastic Models of CDMA Wireless Networks, Chair: Philip Fleming, Authors: Kimberly Wasserman; R. Leelahakriengkrai and R. Agrawal; Philip J. Fleming, Burton Simon and Aleksandr Stolyar; Phil Fleming and Rajesh Pazhyannur 4. Statistical Issues in Inventory Control Chair: Martin Puterman Authors: Jeffrey I. McGill and A. J. (Hamish) Taylor; Stephen A. Smith; Yossi Aviv and Awi Federgruen; Xiaomei Ding and Martin L. Puterman 5. Queueing Models: Performance Evaluation and Control Chair: Hong Chen Authors: Xiaoming Liu, Liming Liu, and David D. Yao; Tuell Green and Shaler Stidham, Jr.; Hong Chen and Yiyuan Zhao; Jinfa Chen, David D. Yao, Shaohui Zheng 6. Markov Decision Processes and Games Chair: Eugene Feinberg Authors: Alex Z.-Z. Lin, James C. Bean, and Chelsea C. White, III; Ulrich Rieder; P. Secchi and W. Sudderth; Eugene A. Feinberg 7. Markov Decision Processes: Analysis and Computation Chair: Bernard Lamond, Authors: Linn I. Sennott; Matthew J. Sobel; T.W. Archibald, C. Buchanan, K.I.M. McKinnon, L.C. Thomas; Abdeslem Boukhtouta and Bernard F. Lamond 8. Software Development in Stochastic Modeling. Chair: Donald Gross and Carl Harris Authors: M. L. Chaudhry; Edward P. C. Kao; Winfried K. Grassmann; Carl M. Harris and Donald Gross; 9. Theory and Applications of Correlated Stochastic Systems Chair: Susan Xu Authors: Thomas (Yew Sing) Lee; Rhonda Righter and Zhen Liu; Susan H. Xu and Haijun Li; Haijun Li and Susan H. Xu; Xiuli Chao 10. Queueing Chair: Hayriye Ayhan Authors: Douglas Down; Tava Lennon Olsen; Chris Lott and Demosthenis Teneketzis; H. Ayhan, F. Baccelli 11. Applied Probability Models in Production Chair: Liming Liu Authors: Hong Chen and Hanqin Zhang; Houmin Yan and John Liu; Siu Wa Ng and Liming Liu; Liming Liu, Xue-Ming Yuan, and John J. Liu 12. Applied Probability Models in Communication Networks Chair: N. Gautam Authors: Les D. Servi; Attahiru S. Alfa, Wei Li; Christopher M. Rump; Natarajan Gautam -------------------------------------------------------------------- LIST OF SPONSORED SESSIONS FOR CINCINNATI -------------------------------------------------------------------- The Applied Probability Cluster for Spring 1999 InfORMS Cincinnati, to be held May 2-5, 1999, has now been organized. The cluster organizer is Tava Lennon Olsen from the Department of Industrial & Operations Engineering at the University of Michigan, Ann Arbor. The sessions are as follows. SESSION 1 TITLE: Advances in the Analysis of Polling Systems SESSION CHAIR: Diwakar Gupta, Michael G. DeGroote School of Business, McMaster University SESSION 2 TITLE: Control of networks under differentiated QoS SESSION CHAIR: Constantinos Maglaras Grad. Sch. of Business, Columbia Univ. SESSION 3 TITLE: Queues: Theoretical and Applied Approaches SESSION CHAIR: Sudha Jain Department of Statistics, University of Toronto SESSION 4: Tutorial: 1998 Erlang Prize Recipients, J. Dai & X. Chao SESSION CHAIR: Tava Olsen IOE Dept., University of Michigan SESSION 5 TITLE Reliability and Maintenance Models SESSION CHAIR: Hayriye Ayhan School of Industrial and Systems Engineering Georgia Institute of Technology SESSION 6 TITLE: Adaptive Stochastic Models SESSION CHAIR: Elmira Popova Department of Mechanical Engineering, The University of Texas at Austin SESSION 7 TITLE: Forecasting and Decision Making over Time SESSION CHAIR: Sarah Ryan Industrial and Management Systems Engineering University of Nebraska-Lincoln SESSION 8 TITLE: Applied Probability in Finance SESSION CHAIR: Steve Kou Dept. of IEOR, Columbia University SESSION 9 TITLE: Applied Probability in Networks SESSION CHAIR: Kimberly Wasserman EECS Department, University of Michigan SESSION 10 TITLE: Current Issues in Call Center Research SESSION CHAIRS: Martin Puterman, Faculty of Commerce and Business Administration University of British Columbia Martin Reiman, Bell Labs, Lucent Technologies -------------------------------------------------------------------- CALL FOR SESSION CHAIRS FOR PHILADELPHIA -------------------------------------------------------------------- CALL FOR SESSION CHAIRS Fall 1999 INFORMS Meeting : Philadelphia, November 7-10, 1999 Dear Colleagues in Applied Probability: You are cordially invited to serve as a Session Chair for the FALL 1999 INFORMS meeting in Philadelphia (November 7-10, 1999). If you would like to organize an Applied Probability Cluster Session, please send an email to ngautam@psu.edu by November 1, 1998. The other deadlines for Session Chairs are December 1, 1998 : Submit a list of speakers (usually 4) January 15, 1998 : Submit abstracts of all session talks We sincerely appreciate all your efforts to make INFORMS-Philadelphia a fabulous conference for all of us. Sincerely, Gautam ------------------------------------------------------------------------ N. Gautam Dept. of Industrial and Manuf. Engr, Phone : 814-865-1239 The Pennsylvania State University, Fax : 814-863-4745 207 Hammond Building, Email : ngautam@psu.edu University Park, PA 16802 -------------------------------------------------------------------- CALL FOR PAPERS: THE 38TH IEEE CONFERENCE ON DECISION AND CONTROL -------------------------------------------------------------------- CALL FOR PAPERS THE 38th IEEE CONFERENCE ON DECISION AND CONTROL Crowne Plaza Hotel, 100 N. 1st Street, Phoenix, Arizona December 5-10, 1999 The IEEE Conference on Decision and Control (CDC) is the annual meeting of the IEEE Control Systems Society (CSS), conducted in cooperation with the Society for Industrial and Applied Mathematics (SIAM) and The Institute for Operations Research and the Management Sciences (INFORMS). The thirty-eighth CDC will begin on Tuesday and run through Friday, December 7-10, 1999, with tutorial workshops preceding the conference on Sunday and Monday, December 5-6. The Conference General Chairman is Edward W. Kamen of the Georgia Institute of Technology (e-mail: kamen@ee.gatech.edu), and the Program Chairman is Christos G. Cassandras of Boston University (e-mail: cgc@bu.edu). The venue is the Crowne Plaza Hotel in downtown Phoenix, Arizona. The hotel is a short walk from restaurants, shops, entertainment spots, and the America West Arena where both professional basketball and ice hockey are played. The warm and sunny climate is ideal for numerous outdoor activities such as golf, horseback riding, hiking; and for one-day or two-day trips to the Grand Canyon, Sedona and the Red Rock Country, and Mexico. Information on the conference and guided tours is available at the CDC99 web site (http://www.ee.gatech.edu/conferences/cdc99). CALL FOR CONTRIBUTED PAPERS AND INVITED SESSIONS The IEEE CDC will include both contributed and invited sessions, and full Proceedings will be published. Contributed papers are solicited in all aspects of the theory and applications of systems, including decision-making, control, adaptation, optimization, industrial automation, and manufacturing. Invited sessions are also solicited in new developments in these and related areas. All submissions are due 1 MARCH 1999. CONTRIBUTED PAPERS: The Program Committee is soliciting both Regular and Short contributed papers for presentation at the conference and publication in the Proceedings. Regular papers describe completed work in some detail. Short papers describe important recent or preliminary results that require limited length. Short papers may be submitted for review as a 4-6 page detailed summary, with references. INVITED SESSIONS: The Program Committee is soliciting proposals for invited sessions. Cohesive sessions focusing on new or emerging topics in the above-listed areas are particularly encouraged, and will have priority over those of a classical or mainstream flavor. Session organizers should submit five (5) copies of proposals for invited sessions for review to the Program Vice-Chair for Invited Sessions, Kevin M. Passino of the Ohio State University (e-mail: passino@ee.eng.ohio-state.edu). All invited session proposals must be accompanied by information for all the proposed papers and mailed together as a group. Organizers must contact the Vice-Chair for Invited Sessions by 1 FEBRUARY 1999 stating their intent to submit a proposal and the accompanying papers. Proposals for invited sessions should contain the names, affiliations, and complete mailing addresses of the session organizer(s), chairperson, co-chairperson, and all authors (with corresponding authors identified) along with a list of paper titles. The organizers must include in the proposal: (1) a clear statement of the topic and purpose of the session and (2) a description of how the papers form a cohesive, well-integrated exploration of the session topic. Detailed extended summaries, following the submission guidelines below and consisting of a minimum of 2000 words, covering all contributions of the paper in sufficient depth to permit an informed review, must be included for each paper. The initial paper may be a tutorial or survey that can be allotted twice the usual time for presentation. The usual number of papers in a session is six. Organizers will be contacted near 1 June 1999 concerning the tentative disposition of their session. SUBMISSION GUIDELINES AND ACCEPTANCE INFORMATION Five copies are required for each submitted regular paper, short paper extended summary, or invited session paper extended summary, and each submission must be accompanied by the Manuscript Submission Form (MSF). Addresses, deadlines, and other submission requirements are given on that form. A procedure for submitting manuscripts electronically is being developed. Instructions for electronic submission, and the MSF, will be posted on the CDC99 web site by October 1998. Final selection of papers and invited sessions will be announced in July 1999. Authors of all accepted papers will be provided instructions for preparation of manuscripts for the Proceedings. Authors should limit their manuscripts to six Proceedings pages (approx. 6000 words) for regular papers or two proceeding pages (approx. 2000 words) for short papers. There will be a mandatory page charge for each additional page. Manuscripts must arrive at the Printer by 16 AUGUST 1999. Authors of accepted papers are expected to attend the CDC to present their work. -------------------------------------------------------------------- CONFERENCE LIST FOR FALL 1998 -------------------------------------------------------------------- Submitted by Brian J. Melloy October 4 - 7 6th Annual Society for Maintenance & Reliability Professionals (SMRP) Conference, held in St. Louis, Missouri. For more information, call 800-950-7354, 312-321-5190, or visit on the Web. October 5 - 7 Simulation of Manufacturing Systems, held in Detroit, Michigan. For more information, call Averill Law at 520-795-6265, fax 520-795-6302, e-mail averill@ix.netcom.com, or web . December 13 - 16 Winter Simulation Conference (WSC) '98, held in Washington, DC. For more information, call Talley Management Group Inc. at 609-423-7222, e-mail meetings@tmg.smarthub.com, or web -------------------------------------------------------------------- INTRODUCING EURANDOM -------------------------------------------------------------------- EURANDOM is a European research institute for the study of stochastic phenomena. Many real life systems and processes in our society are nondeterministic by nature. Mathematical modelling, very useful in analysing, understanding and possibly optimizing such systems, requires therefore the application of 'nondeterministic' mathematics: probability theory, mathematical statistics and the stochastic part of operations research. EURANDOM will carry out research in this area of mathematics and its applications. In short EURANDOM will be, as the name suggests, a EUropean institute for the study of RANDOMness. Formally it is an acronym for: 'European Unit for Research and Analysis of Non-Deterministic Operational Models'. EURANDOM is funded by grants from the government of The Netherlands, Eindhoven University of Technology, The Netherlands Organization for Scientific Research, the European Community (Stimulus) and supported by the national research organizations in several European countries. In particular CNRS (France), DFG (Germany) and FWO (Vlaanderen, Belgium) gave specific commitments. The concept of such a research institute is also supported strongly by industrial and financial institutions. At the moment research projects are being formulated in the following five areas: Complex stochastic models Application of statistics Financial Stochastics Interacting Stochastic Systems Stochastic Networks. EURANDOM is located on the campus of Eindhoven University of Technology in The Netherlands. Its coordinates are EURANDOM, P.O. Box 513, 5600MB Eindhoven, The Netherlands Phone: +31 40 247 8100 Fax: +31 40 247 8190 Email: all@eurandom.tue.nl WWW: http://www.eurandom.tue.nl Scientific director: Willem R. van Zwet Managing director: Wim J.M. Senden -------------------------------------------------------------------- REPORT: EUROPEAN ROUND TABLE ON MAINTENANCE MANAGEMENT (PART 2) -------------------------------------------------------------------- Submitted by Maria F. Ramalhoto. (Editor's note: following Dr. Ramalhoto's account of the Round Table's objectives in our previous issue APN 18(2), here's her report on the Round Tables's discussions and findings.) Maintenance Management: Research & Development Participation Erhan Cinlar, Princeton University, USA; Nozer Singpurwalla, George Washington University, USA; Joel Nachlas, Virginia Tech. USA; Tony Christer, University of Salford, UK; Maria Ramalhoto, IST-Technical University of Lisbon, Portugal; Frank Van der Duyn Schouten, Tilburg University, Holland; Rommert Dekker, Erasmus University, Holland; Martin Newby, City University, UK; Elart von Collani, Wurzburg University, Germany; Shunji Osaki, Hiroshima University, Japan; Ton de Kok, Eindhoven University of Technology, Holland; Roger Cooke, Technical University of Delft, Holland; Stephan Vanneste, Shell Research & Technology Centre, Holland; Bahadir Inozu, University of New Orleans, USA; Attila Csenki, University of Bradford, UK; Wenbin Wang, University of Salford, UK; Xishen Jin, University of Salford, UK; Dieter Girmes, University College London, UK; Harrie de Haas, IVOF, Holland; Jussi Vaurio, Imatran Voima Oy, Finland; Khairy Kobbacy, University of Salford, UK; Raymond Plasmeijer, Erasmus University, Holland; Dave Percy, University of Salford, UK; Jean-François Rit, Electricite de France, France; Antonio Gomez-Corral, IST-Technical University of Lisbon, Portugal (visiting from Complutense University of Madrid, Spain); Amina Bousfiha, University of Technology of Compiegne, France, and the Vice-President of the Portuguese Association of Industrial Maintenance, Portugal. Chaired by Tony Christer and Maria Ramalhoro. In order to initiate the discussion Rommert Dekker, Frank Van der Duyn Schouten, Martin Newby, John Sharp, Joel Nachlas and Tony Christer, were asked to briefly address the questions and topics presented in the Final Program. as well as other additional ones put forward during the Workshop, by the participants. The discussion then followed with Erhan Cinlar, Nozer Singpurwalla, Elart von Collani, Roger Cooke, Ton de Kok, Bahadir Inozu, Maria Ramalhoto, Stephan Vanneste and Jussi Vaurio. Summary of Conclusions of the Round Table discussion on the Gap between Theory and Practice. The Gap It was accepted that a gap exists between the theory and the application of maintenance modelling for decision-making. This statement is, of course, from the perspective of the modellers, who recognise the gap, and not from the perspective of industrial managers, since very few are aware of modelling concepts for decision-making or the practice of modelling for maintenance decisions and, therefore, cannot have an informal opinion. This as we comment later, is part of the problem. The evidence for the existence of the modelling gap is that of the many hundreds of papers published on maintenance modelling, very few address actual situations, and even fewer present validated solutions to recognized problems. Reason for the Gap There are a number of factors considered to be contributing to the existence and even growth of the modelling gap. These include : (i) The real problems are tough to define and usually difficult to analyse and model. (ii) Data is often scattered, incomplete, and lacking in accuracy. (iii) Close collaboration between the modeller and the industrial problem owners is required. This is unlikely to be forthcoming if the organisation does not see a direct and often immediate benefit to themselves, or confidence in the ability to deliver without causing damage in some form. Case experience is helpful in establishing trust and willingness to co-operate, but this only works for the experienced. Thus a chicken and egg scenario exists. This particular problem is not restricted to maintenance, but is particularly acute here because of a lack of recognition within industry of perceived value in maintenance modelling. (iv) There exists a need to influence the culture and attitude towards maintenance within industry, and thereby to require a more capable and better educated manager. Numerous cases exist where a company spends 200,000 dollars on a marketing promotion, will expect to spend 10 - 15 % further in researching its effectiveness. The same company can spend many millions of dollars on maintenance, but spend nothing investigating or researching the effectiveness of the spend. An airforce example was quoted where 15 % of the budget goes on acquisition, 15 % on gasoline, 60 % on maintenance, and 10 % on the rest. Although maintenance costs are four times acquisition costs almost all the research funds go to acquisition problems. (v) Engineers are poorly trained in either the concepts or the use of quantitative methods. Only half a dozen universities address the topic of maintenance on any course either at undergraduate level or postgraduate level, and very few of these address modelling for decision-making. A recent and well funded new course at Manchester University in the general area of maintenance concentrates on the technology of maintenance, and appears to have a very established and traditional content, which does not inform students of the issues being addressed here. Graduates from such course cannot be expected to change the current culture or close the gap. (vi) The system of publication within learned academic journals is academically incestuous since only academics tend to read them. Pressure to publish results in a number of academics developing modelling methods to address conceptual problems with technical interest, but often little immediately obvious relevance. The argument that "though not all the modelling will be realistic, industry can select that which it wishes to use" is not infrequently heard, but is at best rather unworldly. First, industry is insufficiently informed to judge, and secondly they have better things to do with limited time to scan what is seen to be irrelevant material. One could argue that if the maintenance industry understood modelling and used it in maintenance decision-making, they would create the time for more creative thinking through more efficient practice. Closing the Gap (i) Academics along with industrial managers and engineers are, subject to their professional and financial constraints, free spirits, and any movement towards "closing the gap" will need to be generated predominantly by self interest. (ii) Academics generally wish their modelling work to be of influence. Industry will be influenced when the modelling work is seen as clearly addressing recognised problems. (iii) There is a need for academics to move towards industry to identify and solve actual problems. Student placement schemes are a considerable aid here, and collaboration is natural since industry is the laboratory for academic modellers of industrial processes. (iv) Only when a large number of problems have been solved can a more academic overview be established to provide models that apply (in the eyes of industry) to general classes of problems. (v) When collaborating with industry in a modelling exercise, contact is necessary with senior management, with joint ownership by the nominated manager and modeller. (iv) Modellers need to address especially the issues seen as important by industry. (vii) There is no coherent postgraduate training in maintenance addressing maintenance modelling. Students seldom meet the topic of maintenance at undergraduate level or postgraduate level, and seldom meet modelling for decision-making. Post experience courses are required. (viii) In addition to courses, case study presentations by modellers should be given to engineers and maintenance engineering companies. Engineers usually understand their problems well, and if cases are good ones, the concepts of modelling and the benefits can be communicated. This will be a slow process, but hopefully will aid the development of a culture change and the beginning of a "market" for modelling. However, given the current state of training and knowledge, industrial engineers will need considerable assistance from modellers to model their own plant. (ix) Teaching modelling to engineering students and, particularly maintenance modelling, will be a complex task because of the trend to reduce the expected level of quantitative knowledge and ability required of engineers. Unless they can understand elementary statistical and mathematical concepts, their potential to embrace modelling must be limited. This issue cannot be addressed at any but the most fundamental level with university and engineering bodies. (x) It is suggested workshops/courses/conferences be promoted jointly by Engineering Societies and OR/MS groups for post experienced training in the area of maintenance modelling. The requirement is to convey the notion of what modelling is, how it can work, present case examples, motivate the engineer to think in terms of modelling for decision-making in maintenance and to have them recognise that modelling development requires them as an equal partner in the modelling process with the modeller. Other Issue and Emerging Problem Areas (i) The quality revolution has provided dramatic gains in manufacturing efficiency and industrial competitiveness. It is felt by some that the gains available through further enhanced product and process quality activities have almost been fully realised, with little remaining room for improvement in contrast, the competitive advantage and production efficiency attachable through more effective maintenance decision-making and planning is substantial. It is in this area that the next major gains will likely be found. (ii) When modelling a process, there is surprisingly little difference to be found between the large and small companies. In both cases, data is likely to be scattered, incomplete, and of suspect quality (though the data and its significance can be better understood in the smaller organisation). In any modelling exercise, the same relatively modest size of company employers will actually liaise and interact with the modeller, and the same level of dependency will be placed upon the competence and seniority of the industrial contacts. Size of organisation is not a good measure of the intellectual ability of the engineering managers employed, or the ability to produce quality and validated decision-making models. (iii) In data deficient situations, subjective assessment techniques can be very useful. Indeed, even when objective data exists, a subjective exercise can motivate the necessary collaboration for a full modelling exercise. (iv) Health and safety are attracting modelling attention, with particular attention being given to human factors. In part, the solution here lies in there being economic interests in ensuring health and safety. A recent example is the dramatic reduction in serious and fatal accidents in the building industry in Hong Kong. When Hong Kong housing announced no tender would be considered from any contractors without an adequate safety record, practices changed very quickly. -------------------------------------------------------------------- NEW JOURNAL: METHODOLOGY AND COMPUTING IN APPLIED PROBABILITY -------------------------------------------------------------------- A New Journal in Applied Probability Methodology and Computing in Applied Probability Commencing with publication in 1999 Kluwer Academic Publishers plans an international quarterly journal to be be edited by Joseph Glaz from the the University of Connecticut. The aim of the journal is to publish high quality research and review articles that emphasize methodology and computing. Articles presenting theoretical results related to applications and including some computing are wellcome. It is the intent of the editorial board to complete the review of all articles in 4 - 6 month from the time of submission. For instructions for authors and where to submit manuscripts for MCAP please contact Karen Cullen or Joseph Glaz -------------------------------------------------------------------- NEW BOOK ANNONCEMENT -------------------------------------------------------------------- Performance Modelling with Deterministic and Stochastic Petri Nets, 405 pages, hard cover, bundled with DSPNexpress Software on CD-ROM, John Wiley & Sons 1998. Christoph Lindemann, University of Dortmund, Germany (formerly with GMD Research Institute FIRST, Berlin) Web site of the book http://www.first.gmd.de/DSPNexpress/ Book description (from the cover) The first to cover the analysis and application of deterministic and stochastic Petri nets, this book provides comprehensive and fully up-to-date coverage of the fundamental techniques and algorithms. These algorithms are also immediately applicable for the analysis of other discrete-event stochastic systems with an underlying stochastic process that can be represented as a generalised semi-Markov process with exponential and deterministic events. Recent developments in the field have resulted in the need for a text on the subject. This book fills that gap and provides coverage of methodological results on the numerical analysis of deterministic and stochastic Petri nets and their application to performance modelling in parallel computer architecture design. Throughout, the emphasis is on the exposition of an intuitive explanation for the mathematical results rather than rigorous mathematical proof. Consequently, the book is ideal for computer scientists, applied mathematicians and electrical engineers, who are interested in performance analysis of computer systems and communication networks. For readers new to the subject, an introduction to performance modelling is provided, together with appendices containing a primer to applied probability and common probability distributions. The book is accompanied by a CD-ROM containing the software package DSPNexpress for several platforms and specification files of a variety of deterministic and stochastic Petri net models. ISBN 0 471 97646 6 List Price: 70,00 How to order the book with the DSPNexpress software http://www.wiley.co.uk/ordering/ -------------------------------------------------------------------- SOFTWARE ANNONCEMENT: QUEUEING NETWORK ANALYSIS TOOL -------------------------------------------------------------------- QNAT: Queueing Network Analysis Tool URL: http://www.iitk.ernet.in/pub2/ee/networks/qnat/ Authors: D. Manjunath (dmanju@iitk.ernet.in) D.M. Bhaskar Hema Tahilramani (hema@iitk.ernet.in) M.N. Umesh (umesh@sasi.com) Sanjay K. Bose (skb@iitk.ernet.in) We are now distributing an updated beta version of our QNAT package for analyzing queueing networks through a PC-Windows GUI. The software works both on Windows 3.1 and Win95 platforms and uses Mathematica Ver 2.2 (with Mathlink) as the computing engine. You must have Mathematica appropriately installed in your PC to get QNAT to work. We are currently trying to get the software to work using the new version of Mathematica (Version 3). If you are using this version of Mathematica and would still like to use the package, please let us know (skb@iitk.ernet.in). In that case, we will inform you when we finally get it working with this version. Here is a quick list of the types of queueing networks that can be solved using QNAT - 1. Networks of Multi-server Infinite Capacity Queues with Multiple Job Classes - Open Networks - Closed Networks - Mixed Networks, with some job classes closed and others open 2. Networks of Multi-server Infinite Capacity Queues with a Single Job Class - Open Networks without Fork-Join Nodes - Open Networks with Fork-Join Nodes without Synchronization - Closed Networks without Fork-Join Nodes - Closed Networks with Fork-Join Nodes with Synchronization - Closed Networks with Fork-Join Nodes without Synchronization 3. Networks of Multi-Server Finite Capacity Queues with a Single Job Class (some or all nodes may be of finite/infinite capacity) -Open Networks with Transfer Blocking -Open Networks with Repetitive Service, Random Destination Blocking -Open Networks with Repetitive Service, Fixed Destination Blocking -Open Networks with Rejection Blocking -Closed Networks with Transfer Blocking -Closed Networks with Repetitive Service, Random Destination Blocking -Closed Networks with Repetitive Service, Fixed Destination Blocking Our distribution includes documentation on the analytical techniques that have been used. This has been provided through a Microsoft Word file that is a part of this distribution. The documentation provides a summary of the actual algorithm used by each solution technique in QNAT and examples of networks solved using this technique in QNAT. These examples may be found in the samples distributed with QNAT. This beta version of QNAT is being distributed free over the Internet. Please contact Prof. Sanjay K. Bose (skb@iitk.ernet.in) for current information on the sites from where this beta version of QNAT may be downloaded. -------------------------------------------------------------------- OPENING: POST-DOCTORAL POSITION -------------------------------------------------------------------- 2 year Post-Doctoral Position at AT&T Labs, Florham Park, NJ. MINC: Multicast-based Inference of Network-internal Characteristics This is a announcement of a two year project starting September 1998 to investigate the use of end-to-end multicast traffic measurements for the purpose of inferring link-level performance characteristics such as link loss and link delay statistics. The project is funded in part by DARPA under BAA 98-02, Next Generation Internet: Network Engineering. The project involves researchers at Univ. of Massachusetts (D. Towsley, J. Kurose, J. Horowitz), AT&T Labs-Research (N. Duffield, R. Caceres), and Lawrence Berkeley National Laboratory (V. Paxson). We are looking for a post-doctoral researcher for the two-year period. This person should either have a solid background in statistics and estimation theory and some knowledge of networking or a solid networking background along with some knowledge of statistics and estimation theory. The postdoc will be located at AT&T Labs-Research in Florham Park, New Jersey. Further details of the MINC project can be found at http://www.research.att.com/~duffield/minc/ Expressions of interest or questions should be addressed to Nick Duffield (duffield@research.att.com) or Don Towsley (towsley@cs.umass.edu) -------------------------------------------------------------------- OPENING: CHAIR IN MANAGEMENT SCIENCE/OPERATIONS RESEARCH -------------------------------------------------------------------- The University of Canterbury invites applications for a CHAIR IN MANAGEMENT SCIENCE/OPERATIONS RESEARCH, made available by the retirement of Professor HG Daellenbach. The appointee will be expected to provide academic leadership by encouraging excellence in research and teaching, including through example. Supervision of doctoral students and a strong ongoing record of publications in top internationally recognised journals in MS/OR and related areas will be considered as proof of academic leadership. Good teaching skills are expected. In order to further strengthen the MS/OR group, the appointee is expected to have an international reputation for research in an analytic area of MS/OR, with strong preference given to stochastic areas, such as Decision Analysis, Simulation, Systems Dynamics, Queueing, Forecasting, or Problem Structuring Methods. Candidates should also have demonstrated interest and competence in applied research of value to business, industry, or the public sector. The appointee will contribute towards teaching in one or more of the undergraduate courses, and at the graduate level in her/his area of expertise, including PhD and Masters research supervision. The MS/OR groups consists of 6 operations researchers and 3 staff in production/manufacturing management. It offers B.Sc. and B.Com. degrees in MS/OR and in Production Management, one and two-year graduate programs in Operations Research, and PhDs by supervision. There is a strong energy modelling group with an international standing. Consulting work is encouraged. The teaching load is approximately 100 contact hours per year. This usually allows ample time for periods of uninterrupted research during a substantial portion of the year. The University also has generous sabbatical leave conditions, providing for a full-year at full pay after six years of service, plus a travel allowance. For shorter periods of service, shorter leave conditions apply. The department normally finances an annual trip to the Northern Hemisphere for presentation at a conference. The current salary for Professors is in the range of NZ$86,500 to 106,500 per year, which is in the top 2-3% of salaries in New Zealand. Enquiries of an academic nature may be made to Prof. HG Daellenbach, Department of Management, Tel. (64 3) 364 2987 ext. 8613, Fax (64 3) 364 2020, or e-mail: . For details on the Department of Management, the Management Science group within the department, staff and their research interests, and the current offerings of courses, consult our web page at . Conditions for appointment can be obtained from the Human Resources Registrar (e-mail: ) or the address below. Short-listed candidates will be invited to visit the University and give presentations on their current and past research. Applications close on October 30, 1998. Prospective candidates are requested to submit their application, including a full CV, samples of recent publications, and the names of three academic refereees to the Human Resource Registrar, University of Canterbury. Private Bag 4800, Christchurch, New Zealand, quoting Position No. MG17. ****************************************************************** Prof. H. G. Daellenbach, Head Department of Management, University of Canterbury Private Bag 4800, Christchurch, New Zealand Phone: 64 3 3642987 X8613; FAX: 64 3 3642020 e-mail: h.daellenbach@mang.canterbury.ac.nz -------------------------------------------------------------------- EDITOR'S COORDINATES -------------------------------------------------------------------- Please note that, due to a host server upgrade, the subdirectory for the Applied Probability Section's web page has been temporarily redirected (www3 instead of www). The redirection is automatic but, as a side effect, it defeats the "Go Back" menu on the web browser. Things should be back to normal in January. I look forward to receiving your input for the Applied Probability Newsletter and the Web page. Bernard F. Lamond, Ph.D. Department of Operations and Decision Systems School of Business Administration Laval University Quebec, PQ CANADA G1K 7P4 Email: URL: Telephone: 418-656-2131x5472 Facsimile: 418-656-2624 Applied Probability Home Page: ====================================================================