Depto. Investigaci

1
Depto. Investigación Operativa,Computer Science
Institute,Engineer Faculty Universidad de la
República, Uruguay
March 2005
2
Staff of the group

• Assistants
• Celina Gutiérrez
• Antonio Mauttone
• Alfredo Olivera
• Martín González
• Pablo Rodríguez Bocca
• Eight PhD and MsC students
• Students working on their degree dissertations.

• Professors
• Héctor Cancela
• María E. Urquhart
• Omar Viera
• Carlos Testuri
• Franco Robledo
• Graciela Ferreira
• Daniel Meerhoff

3
Courses

• Introduction to Operation Research.
• Extensions to Linear Programming and Integer
  Programming.
• Metaheuristics solving Combinatorials Problems.
• Optimization under uncertainty.
• Discrete Event Simulation.
• E-commerce and OR.

4
Main research areas and methodologies

• Main application areas
• Transportation,
• Logistics,
• Communications.
• Cooperation areas
• High performance Computing.
• Other areas of interest
• Finance,
• Electricity networks,
• Re-engineering.

• Methodologies
• Optimization
• Heuristics, meta-heuristics.
• Mathematical programming,
• in particular
• stochastic programming and dynamic
  programming.
• Simulation
• Discrete event simulation.
• Monte Carlo methods,
• Variance Reduction methods.

5
Transportation

• Vehicle routing
• Cluster first, route second" methods (Omar
  Viera, Libertad Tansini, Daniel Giosa, Marita
  Urquhart)
• Routing heuristics.
• Assignment heuristics, Genetic Algorithms.
• Metaheuristics
• Ant Systems (Hector Cancela, Omar Viera).
• Adaptive memories (Alfredo Olivera, Omar Viera).

6
Article

• Adaptative Memory Programming for the Vehicle
  Routing Problem With Multiple Trips.
• A. Olivera, O. Viera (2005)
• Difficulty each vehicle may perform several
  routes in the same planning period, it implies
  the design of a set of routes and the
  assignmentnt of those routes to the available
  vehicles.
• Heuristic based on AMP (Rochat and Taillard).
• Tested over the 104 benchmark problems proposed
  by Taillard et.al.

7
Transportation

• Urban public transport (H.Cancela, M.E. Urquhart,
  A. Mauttone. O. Viera)
• Bus line simulation.
• Origin-destination-time demand matrix estimation.
  
• Bus line design.
• Intelligent Transport Systems.
• Vertical transport (H. Cancela, O. Viera)
• Capacity estimation.
• Simulation and comparison of scheduling
  algorithms.

8
Extended Abstract and Posters

• Bus lines design in urban public transport 
• A.Mauttone, M.E. Urquhart (ICIL Poster-2005).
• GRASP multiobjetivo para el diseño de rutas y
  frecuencias en transporte colectivo público
  (CLAIO 2004)
• This work is about the design of bus lines
  (routes and frequencies) in urban public
  transport systems. It is part of a long reach
  project, whose objective is to develop a tool
  that supports the transit network design planning
  in a strategic context.
• Characteristics
• High combinatorial complexity.
• Multi-objective nature objectives of users and
  operators expressed in different units, and
  usually contradictory (an improvement in one
  objective implies a detriment in the other one).
• Mixed variables discrete variables represent
  the sequences of nodes in each bus route, and
  continuous variables represent the bus
  frequencies.
• Real Case study with real data

9
Communication networks

• Wide Area Network topology design (H. Cancela).
• Steiner trees, 2-connected networks, Generalized
  Steiner problems in graphs
• Ant Systems
• GRASP (F. Robledo)
• Genetic Algorithms
• Other heuristics.
• Topology design with uncertain demands
  (stochastic programming) (C. Testuri).

10
Networks

• Reliability models (H.Cancela)
• Classical models evaluation using variance
  reduction techniques. ( and G. Rubino)
• Models with distance restrictions. (and L.
  Petingi)
• Optimization - GRASP, Genetic Algorithms.
• Peer-to-peer network models.
• Performance evaluation of computer clusters
• Petri networks.
• Markov chains, simulation.

11
Article

• Network design with node connectivity
  constraints H. Cancela, F. Robledo, G.
  Rubino.(2004)
• Problem to find the least costnetwork topology
  that satisfies some node-disjoint path connection
  requirements
• Modeled as a GSPNC (Generalized Steiner Problem
  with Node-Connectivity Constraints)
• Solved with an algorithm based on GRASP.
• Present experimental results, test problems with
  more than 100 nodes, obtained from public
  libraries.

12
Article

• A GRASP algorithm with RNN based local search
  for designing a WAN acces network H. Cancela, F.
  Robledo, G. Rubino.(2004)
• Problem to design the access network topology
  of a Wide Area Network
• Modeled as a variant of the Steiner Problem in
  graphs.
• Solved with an algorithm based on GRASP that
  employs Random Neural Networks as local search
  heuristic.
• Present experimental results, test problems
  obtained from the SteinLib .

13
Article

• Evolutionary algorithms applied to reliable
  communication network design
• S. Nesmachnow, H. Cancela, E. Alba.(2004)
• Problem to design a reliable communication
  network topology
• Modeled as a Generalized Steiner Problem in
  graphs (to design a minimum cost subnetwork
  verifying a set of prefixed minimum connection
  requirements for pairs of terminal nodes).
• Solved with a set of pure and hybrids
  Evolutionary algorithms (GA, CHC, GASA,
  Parallel)
• Present a comparative study of them with a
  random test suite of three .

14
Technical reports

• Diseño de redes de Comunicaciones con
  Incertidumbre en la demanda Degree dissertation.
• J. Viera, C. Testuri.(2004)
• Multicommodity network, Sotchastic Programming.
• Motor de metaheurísticas para el problema de
  asignación de salones de la Facultad de
  Ingeniería
• S. Costabel, M. Urquhart (2005). Master
  thesis.
• Timetable problem, Tabu-Search, Genetic
  Algorithm, Ant-systems. Real test-case
  Engineering Faculty.

15
Simulation, other topics

• Discrete event simulation (H.Cancela.
• M.Urquhart).
• Scheduler for DES. L.Murray
• Bus line design. A. Mauttone et.al

16
Projects with industry and public sector

• CoNaProLe truck routing for milk collection.
• IMM, IMR urban public transport studies.
• ANCAP / DUCSA preliminary study of opportunity
  areas for optimization.
• MGAP study of organization of computing center.
• BPS third party services..

17
International cooperation

• INRIA (Institut National de Recherche en
  Informatique) Rennes, Francia Team Planning of
  Architecture and Infrastructre of Networks ,
  joint projects, students co-direction.
  (Communication Networks design)
• City University of New York joint
  projects.(Networks design diameter restriction)
• PROSUL network Discrete optimzation and graphs
  theory, algorithms and applications USP, UFRJ,
  U. Chile, UBA, U. Rosario, etc.
• ALFA network Solving real problems of high
  complexity in combinatorial optimization U.
  Bologna, Avignon, L. de Bruselas, U. Concepción,
  etc. Students co-direction.

18
Events organized

• IV and X ELAVIO (Escuela Latinoamericana de
  Verano de Investigación Operativa) Montevideo
  1997, 2004.
• ICIL 2005 (International Conference on Industrial
  Logistics) Montevideo, february 2005.
• CLAIO 2006 (Congreso Latino Ibero Americano de
  Investigación de Operaciones) Montevideo,
  november 2006.