| Home | | | Contact | | | Program | | | Newsletter | | | Membership | | | OR in Melbourne | | | Links | | | Archive |
2014 Program
The default venue for the monthly lectures is RMIT, Access Grid Room. 8.9.66
Melbourne time
Scheduled Events for 2014
Venue: Monash University, Clayton: Room G12A, Building 26 (Live video Caulfield campus, Building H room 7.84 (level 7))
Time/Date: 12:00PM, November 11, 2014
Program: Seminar
Topic: Perturbed Perspective Cuts in tight Unit Commitment models with Optimal Transmission Switching
Abstract
In this talk we address the issue of solving a Unit Commitment (UC) Problem including the electricity transmission network with Active Switching (AS). The switching operation consists in a dynamic reconÞguration of the electricity network, i.e. opening or closing some lines; this paradigm is named UC with Optimal Transmission Switching (UCOTS). The UCOTS is a novel way to leverage grid controllability, that may reroute the electrical energy in the network to resolve some line congestion and reduce the overall production cost, e.g. by allowing to increase the power output of cheaper units and/or at improving other system performances such as stability. This work provides a tight Mixed Integer Linear Programming (MILP) formulation for the UCOTS. It proposes to approximate the quadratic thermal cost objective function by means of a perspective cuts (PC) piece-wise-linear function and to perturb it in a special way, in order to break some of the symmetries deriving from the OTS variables. We show that, combining these two ingredients and using recent tight UC constraints formulations, optimal and near-optimal solutions can be obtained in reasonable computing time for a custom 81 units-IEEE 118 bus test case.
Bio. Giovanni Felici is currently a Senior Research Scientist at the Institute of System Analysis and Computer Science of the Italian National Research Council (Consiglio Nazionale delle Ricerche, CNR), located in Rome, where he holds a tenured position since 2000. He graduated in Statistics from the University of Rome ÒLa SapienzaÓ, received a Master of Science in Operations Research and Operations Management from the University of Lancaster (United Kingdom), and a Ph.D. in Computer Science from the University of Rome. His current research interests cover the application of mathematical modelling and optimization algorithms to logistics, scheduling, data mining, and bioinformatics.
Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:30PM, Wed September 17, 2014
Program: Lecture by Alysson M. Costa, University of Melbourne
Topic:: The vegetable crop rotation scheduling problem with minimal plot sizes*
Abstract
* Joint work with Lana Mara R. dos Santos, Pedro Munari and Ricardo H. S. Santos.
Crop rotation plays an important role in agricultural production models with sustainability considerations. Commonly associated strategies include the alternation of botanical families in the plots, the use of fallow periods and the inclusion of green manure crops. In this article, we address the problem of scheduling vegetable production in this context. Vegetables crop farmers usually manage a large number of crop species with different planting periods and growing times. These crops present multiple and varied harvesting periods and productivities. The combination of such characteristics makes the generation of good vegetable crop rotation schedules a hard combinatorial task. We approach this problem while considering two additional important practical aspects: standard plot sizes (multiples of a base area) and total area minimisation. We propose an integer programming formulation for this problem and develop a branch-price-and-cut algorithm that includes several performance-enhancing characteristics. Computational experiments over a set of instances based on real-life data are presented.Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:30PM, Wed September 3, 2014
Program: Lecture by Claus Brech, TU Munich and CSIRO
Topic:: A decision support system for scheduling resident training at university hospitals
Abstract
Residents in training gain experience by providing assistance in the operative theatre, particular to their surgical specialty. They must assist on a certain quota of surgeries to attain their training requirement in each subspecialty. A resident training schedule has to determine the number and type of surgeries a resident in training must complete within a given period. This paper proposes a decision support system that aims to assist planners in scheduling resident training. The system is based on a hierarchical planning approach. On the first level an optimization model is used to determine monthly assignment of residents to surgeries with the objective to minimize the overall tardiness of resident training. The second level decides on the assignment of residents to surgeries on a daily basis with the objective to meet the monthly training requirements determined on the first level. The presentation will mainly focus on the second level. We model the dynamic assignment problem as a Markov Decision Process (MDP). The MDP become intractable to exact methods due to the curse of dimensionality. We address this problem by proposing an Approximate Dynamic Programming approach that approximates the value function using a linear architecture. Computational experiments based on real data from a German teaching hospital demonstrate the performance of this approach.Bio
Claus Brech is a PhD student at the Chair of Operations Management, Technical University of Munich. His main research interests are Operations Research, Decision Support Systems and Health Care Operations Management. He completed his master degree in Information Systems at the Technical University of Darmstadt in 2010, majoring in Operations Research and Supply Chain Management. Claus is currently at the CSIRO as a visiting researcher under the supervision of Dr. Andreas Ernst.Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:30PM, Wed August 20, 2014
Program: Lecture by Leorey Marquez, CSIRO
Topic:: Application of Participatory Risk Assessment for Creating Disaster Risk Sensitive Shelter Plans
Abstract
Disasters disproportionally affect the poor, vulnerable, and marginalized. Improving social protection and community-driven development (CDD) interventions and linking these mechanisms to disaster risk management (DRM) increases the effectiveness of DRM programs in contributing to poverty alleviation and sustainable development, while significantly reducing disaster response costs. These interventions are especially relevant to the Philippines which has the third highest disaster risk worldwide, and with climate change, high population density and poverty conditions, the impacts of disasters are expected to increase. This presentation describes the implementation and impact of a CDD project aimed at creating Disaster Risk-Sensitive Shelter Plans (DR-SSP) for nine barangays in the city of Legazpi. The results show the importance of building capacity in civil society organizations and local government, and the effectiveness of the participatory approach in data collection, risk assessment, and disaster planning.Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:30PM, Wed July 16, 2014
Program: Lecture by Dudley Foster, University of Edinburgh
Topic:: Reflections on a Lifetime of OR and its Application in Private Life
Abstract
In this talk, Dudley will reflect on a lifetime's OR experience, covering, in particular: issues of problem definition; the use of scenarios as an aid to decision making under uncertainty; network analysis (CPM / PERT) as a formal technique, but also using 'network thinking' to prioritize work both professionally and personally; Queuing Theory, and the widespread failure to use simple, long established knowledge; Kepner Tregoe methodology, especially Potential Problem Analysis; and basic principles of implementation and client relations, including their use in reducing the trauma associated with managing emotionally laden issues in private life. Each topic will be illustrated with one or two examples from DudleyÕs professional experience and also by more recent examples from his private life and/or the planning of professional activities.The talk will conclude with a brief outline of a voluntary sector project (design and construction of an anelammatic sundial Ð including the writing of user guide at a reading age of 10) carried out for Rotary Club of Leven in a multi - disciplinary team consisting of: a fund raising expert (the project was a condition attached to one source of funds); a mathematician (Dudley); a structural engineer (whose office did the technical drawings and who helped to physically validate the design); a retired banker (who came up with excellent cost - reducing ideas) and a retired primary school teacher (whose expert help was essential in correcting the reading age of the User Guide) .
Bio:
Dudley Foster, B.A (Hons), M Tech (OR) has 40 years of experience in Operations Research in the UK, Australia and New Zealand., During his last 10 years in Australia (1996 to 2006) he was a Director of DNF Decision Sciences Pty Ltd and latterly of NORCA Consulting. In addition to his 14 years as a consultant, his experience includes 18 years in industry (6 years with the NCB in UK and later 12 years with Shell Australia) and 6 years as a tertiary lecturer, during his first 6 years in Australasia.A lot of his work was related to business improvement across the interface between Manufacturing and Marketing. He is also an expert on financial modelling, with a particular focus on the economic evaluation of investment proposals and has developed a number of generic financial models, as well as tailored models covering greenfield projects, acquisitions and divestments. His course Financial Modelling in Excel was delivered many times under the auspices of AIOR, both as a public offering and to in - house groups, and more recently, the course has been translated into Spanish, in collaboration with a Mexican associate, Ricardo Melgar Cruz.
During his last 10 years in Australia, Dudley was a Senior Academic Associate with the Graduate School of Business at Victoria University and the Industrial Adviser to the Financial Modelling [Research] Program. Since retiring and returning to UK, he has worked as a sessional supervisor of projects undertaken by students on the MSc in OR at the University of Edinburgh, where he also attends seminars at the Edinburgh Research Group in Optimization (ERGO), the Edinburgh equivalent of ORSUM at the University Melbourne. A synopsis of the one seminar he delivered to ERGO can be found at http://www.maths.ed.ac.uk/ERGO
Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:30PM, Wed June 18, 2014
Program: Lecture by Xiaodong Li, RMIT
Topic:: Large Scale Global Optimization
Abstract
Many real-world optimization problems involve a large number of decision variables. For example, in shape optimization a large number of shape design variables are often used to represent complex shapes, such as turbine blades, aircraft wings, and heat exchangers. However, existing optimization methods are ill-equipped in dealing with this sort of large scale global optimization (LSGO) problems. A natural approach to tackle LSGO problems is to adopt a divide-and-conquer strategy. A good example is the early work on a cooperative coevolutionary (CC) algorithm by Potter and De Jong (1994), where a problem is decomposed into several subcomponents of smaller sizes, and then each subcomponent is ?cooperatively coevolved? with other subcomponents.In this talk I will provide an overview on the recent development of CC algorithms for LSGO problems, in particular those extended from the original Potter and De Jong?s CC model. One key challenge in applying CC is how to best decompose a problem in a way such that the inter-dependency between subcomponents can be kept at minimum. Another challenge is how to best allocate a fixed computational budget among different subcomponents when there is an imbalance of contributions from these subcomponents. Equally dividing the budget among these subcomponents and optimizing each through a round-robin fashion (as in the classic CC method) may not be a wise strategy, since it can waste lots of computational resource. Many more research questions still remain to be answered. In recent years, several interesting decomposition methods (or variable grouping methods) have been proposed. This talk will briefly survey these methods, and identify their strengths and weakness. The talk will also describe a contribution-based method for better allocating computation among the subcomponents. Finally I will present a newly designed variable grouping method, namely differential grouping, which outperforms those early surveyed decomposition methods.
Bio:
Xiaodong Li received his B.Sc. degree from Xidian University, Xi'an, China, and Ph.D. degree in information science from University of Otago, Dunedin, New Zealand, respectively. Currently, he is an Associate Professor at the School of Computer Science and Information Technology, RMIT University, Melbourne, Australia. His research interests include evolutionary computation, machine learning, complex systems, multiobjective optimization, and swarm intelligence. He serves as an Associate Editor of the IEEE Transactions on Evolutionary Computation, the journal of Swarm Intelligence, and International Journal of Swarm Intelligence Research. He is a founding member and currently a Vice-chair of IEEE CIS Task Force on Swarm Intelligence, and currently a Chair of IEEE CIS Task Force on Large Scale Global Optimization. He was the General Chair of SEAL'08, a Program Co-Chair AI'09, and a Program Co-Chair for IEEE CEC?2012. He is the recipient of the 2013 ACM SIGEVO Impact Award. For further information, please visit his website:
http://goanna.cs.rmit.edu.au/~xiaodong/Related Publications:
Omidvar, M., Li, X., Mei, Y. and Yao, X. (2013), "Cooperative Co-evolution with Differential Grouping for Large Scale Optimization", IEEE Transactions on Evolutionary Computation (accepted on 21 May 2013).
Omidvar, M., Li, X. and Yao, X. (2011), "Smart Use of Computational Resource Based on Contribution for Cooperative Co-evolutionary Algorithms", in Proceeding of Genetic and Evolutionary Computation Conference (GECCO'11), ACM Press, p.1115 - 1122.
Tang, K., Li, X., Suganthan, P.N., Yang, Z. and Weise, T. (2010), "Benchmark Functions for the CEC'2010 Special Session and Competition on Large Scale Global Optimization", Technical Report, Nature Inspired Computation and Applications Laboratory, USTC, China; URL:http://nical.ustc.edu.cn/cec10ss.php.
Yang, Z., Tang, K. and Yao, X. (2008), ?Large scale evolutionary optimization using cooperative coevolution?, Information Sciences, 178:2986?2999, August 2008.
Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:30PM, Wed May 14, 2014
Program: Lecture by John Hooker, Carnegie Mellon University
Topic:: Discrete Optimization with Decision Diagrams.
Abstract
We propose a general branch-and-bound algorithm for discrete optimization in which binary decision diagrams (BDDs) play the role of the traditional linear programming relaxation. Relaxed BDD-based problem representations of bounded size provide bounds and guidance for branching, while restricted BDDs supply a primal heuristic. The BDDs also allow more effective constraint propagation than domain consistency propagators. The problem is modeled as a dynamic programming (DP) recursion, even though it is not solved by DP. This opens the possibility of solving instances too large to load in an integer programming (IP) solver. A novel search scheme branches within relaxed BDDs rather than on values of variables. Preliminary testing shows that a rudimentary BDD-based solver is superior to a leading commercial IP solver for the max stable set problem, the max cut problem on a graph, and the max 2-SAT problem. In addition, we improved the best known solutions and tightest known bounds for several large-scale max cut benchmark instances. This is joint work with David Bergman, Andre Cire and Willem van Hoeve.Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 4:00PM, Wed April 16, 2014
Program: Lecture by Martijn Van der Merwe, RMIT
Topic:: A mixed integer programming approach for asset protection during escaped wildfires.
A reminder of our student focussed seminar at 4pm on April 16, 2014. Note the earlier, student friendly start time.
When: 16:00-17:30 AEST, Wednesday April 16th, 2014
Speaker Live From: RMIT University, City Campus, Room 8.9.66 (Access Grid)
Access Grid Sites: The University of Newcastle, Rooms v205/v206
This is an opportunity for students to socialise, meet industry practitioners and academics, and exchange views and experiences.
Refreshments will be served at both RMIT and Newcastle.
ASOR Melbourne committee members will be present at each location, and of course the invitation to attend extends equally to our non-student members as well. For those in research and industry, this will be a great way to make contact with the future of OR and we encourage you to attend.
The RMIT venue is Room 8.9.66 (the Access Grid room, our usual venue for ASOR events), and the Newcastle venue is Rooms v205/v206 (also the usual Access Grid rooms used for ASOR events) and it is *free*. Indeed -- free ASOR membership for the rest of the year is on offer to all students who attends! (OK, I admit, student membership is free anyway!)
Please note that this event is not restricted to ASOR members, and that we would appreciate if members could circulate this invitation to any students that they may be supervising and working with.
For enquiries, please contact either Cam MacRae [RMIT] ( cameron.macrae{at}rmit.edu.au), or Alan Lee [Newcastle] (alee4{at}uon.edu.au).
Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:00PM, Wed March 19, 2014
AGM
The ASOR Melbourne Annual General Meeting will be held on Wednesday 19 March at 5pm in the Access Grid Room at RMIT (Building 8, Level 9, Room 66).
We will join regional/interstate sites by AGR, and/or possibly by telephone, as requested (prior!) by people at those sites.
ASOR Melbourne chapter covers not only Melbourne, but we also have members in Adelaide, Newcastle, Ballarat, and even Belgium. This slightly strange state of affairs follows on from ASOR's slightly strange structure.
The main function of the AGM is to elect a new committee (for April 2014 to March 2015). Nominations for committee positions are now open. A nomination form will be sent in an email that follows this one.
The committee meets at 4pm on the third Wednesday of each month. The committee's main role is to facilitate the delivery of services to members:
- Seminars
- Workshops (such as the Energy Modelling and Health Systems Modelling workshops)
- Conferences (such as the ASOR National Conference and Recent Advances)
- Website and Newsletter
The committee has a maximum of 12 elected members (additional members can subsequently be co-opted). This includes five "executive" positions:
- Chair (aka President)
- Vice-Chair
- Secretary
- Treasurer
- Student Representative
- Two newly-defined roles/responsibilities:
- SA (Adelaide) Events Liaison
- NSW (Newcastle) Events Liaison
Inferred by the above is that we are seeking at least one NSW and one SA member of the committee.
Regards,
Simon Dunstall
Chair ASOR Melbourne 2013Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 5:30PM, Wed March 19, 2014
Program: Lecture by Melih Ozlen, RMIT
Topic:: Recent developments in multi objective mixed/integer/linear programming
Abstract
It has been a very exciting couple of years for the field of multi objective mixed/integer/linear programming with a renewed interest in the area. New algorithms were developed, either to generate efficient or nondominated sets or optimise functions over these sets by identifying the best solutions. Reasonably large problem instances with more than three objectives were solved for at least certain problem types. Classical algorithms were revisited to improve their performances significantly. Highly parallel computing environments were utilised by algorithms taking advantage of them. Bounds were developed on the number of subproblems that needs to be solved to generate nondominated sets. Open-source implementations of some of these algorithms were made available online.In this talk, I will briefly summarise these exciting years, comment on the current state of research in multi objective mixed/integer/linear programming, and will discuss some future research directions.
Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 6:00PM, Wed February 19, 2014
Program: Lecture by Bruce J. Miller, Principal Consultant Power Systems, WorleyParsons
Topic:: A simulation and analysis of the 5-state power system between Monday 13th to Friday 17th January
Abstract
Bruce will take us through the methods and results of some national electricity system simulation modelling and analysis that has been carried out by WorleyParsons and centred on the heatwave conditions earlier this year.Venue: RMIT Access Grid Room, 8.9.66 (Building 8, level 9, room 66)
Time: 6:00PM, Wed February 26
Program: Lecture by Prof. Sally Brailsford, University of Southampton (UK)
Topic:: Discrete-event simulation is alive and kicking!
Abstract
At the 2010 UK OR Society Simulation Workshop, there was a lively panel discussion entitled "Discrete-event simulation is dead, long live agent-based simulation!" which was subsequently written up as a position paper for the Journal of Simulation. This talk continues that discussion and, to quote Mark Twain, argues that rumours of the death of DES are greatly exaggerated. There has undoubtedly been a recent surge of interest within the mainstream OR community in the use of agent-based modelling, but this talk suggests that many of the cited benefits of ABS can be achieved through the use of a traditional DES approach. These arguments are illustrated by several examples where DES has been used successfully to tackle "ABS-type" problems.
IFORS 2011 Conference
July 10-15, 2011, Melbourne, Australia