Heuristic Optimization

Planning

Dates: November, 14-18
Room: Sala Nerja
Start End Day 1 Day 2 Day 3 Day 4 Day 5
10:00 11:45 · Presentation
· Introduction
· GA (Genetic Algorithms) · EDA · Genetic programming
· High-Performance optimization
· Multiobjetive optimization2)
11:45 12:00 Coffee break
12:00 13:15 · Simple methods · Interactive GA
· Micropopulation GA
· EDA
· Hybrid algorithms
· Dificulty on optimization problems · Other techniques
13:15 14:00 Lunch
14:00 16:00 Practical work:
Simulated Annealing 3)
(Map Colouring)
Practical work:
Genetic Algorithms
(Map Colouring)
Practical work:
EDA
(TSP)
Practical work:
High-Performance Optimization
Free time for finishing practical works
EXAM

Detailed Outline

  • Day 1:
    1. Presentation (1:00): Víctor Robles & Oscar Cubo Medina
    2. Introduction to heuristic optimization (0:30): Oscar Cubo Medina
      1. Simple examples
      2. Real applications
    3. Simple methods (1:30): Oscar Cubo Medina
      1. Local Search
        1. Hill-Climbing
      2. Tabu Search
      3. Simulated Annealing
    4. Practical work - Simulated Annealing: Pilar Herrero & Santiago González
      • Software: C or R
      • Problem: Map Colouring
  • Day 2:
    1. Genetic Algorithms
      1. Genetic Algorithms (2:00): María S. Pérez
      2. Interactive GA (0:30): José M. Peña
      3. Micropopulation GA (0:30): José M. Peña
    2. Practical work - Genetic Algorithms: Antonio LaTorre
      • Software: C (GA-EDA-Lib)
      • Problem: Map Colouring
  • Day 3:
    1. Estimation Distribution Algorithms (1:30): Víctor Robles
    2. Hybrid algorithms (1:30): Víctor Robles
      1. Taxonomy
      2. Lamark and Baldwin approaches
      3. GA-EDA approach
    3. Practical work - EDA: Oscar Cubo Medina
      • Software: C (GA-EDA-Lib)
      • Problem: TSP
      • Extras:
        1. Local Search
        2. GA-EDA approach
  • Day 4:
    1. Genetic programming (1:00): Antonio LaTorre
    2. High-Performance optimization (1:00): José M. Peña
    3. Dificulty on optimization problems (1:00): José M. Peña
    4. Practical work - High-Performance Optimization: Antonio LaTorre
      • Software: C (GA-EDA-Lib)
      • Problem: Typical benchmark (Branin, RCOS, Schefe,…)
      • Extras:
        1. Islands
        2. Comparative between GA, EDA and others approaches
  • Day 5:
    1. Multiobjetive optimization (2:00): Santiago González
    2. Other opimization techniques (1:00): Oscar Cubo Medina
    3. Exam

Evaluation

The evaluation of the course will have two parts:

  • Exam: There will be a small multichoice test. (1 hour aprox.)
  • Practical works: All practical works must be sended by e-mail to ocubo@fi.upm.es before Friday, 25 November.
    The mail must include:
    • Identifier used in the linux computer
    • One small report with comments about the solution of each problem. It may also include personal comments about the whole Athens course.
    • Source code used to implement the proposed solutions (for GA and EDA problems, we could access your solution on the computer)

You could also put your memory and code in the linux account. In this case, the mail must contains your name and the account you use in the course.

Practical works

The course has 3 practical works:

  • Solve the Map-Colouring problem with Simulated Annealing
  • Solve the Map-Colouring problem with Genetic Algorithm
  • Solve the Traveling Salesman Problem with EDAs
    • Compare a comment the diferences with GA (quality of the solution and time)
    • Compare a comment the diferences with GA-EDA approaches (quality of the solution and time)

Other

:!: Don't forget to fill the survey about the course on the Athens Web Page :!:
1) Coordinator
2) A littleExample of multiobjetive optimization
3) Code for Visual Studio 6.0 (it works on VS.NET) with an example of matrix.
 
docencia/cursos/heuristic_optimization.txt · Última modificación: 2012/10/08 17:58 (editor externo)
 
Recent changes RSS feed Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki