print print


Practical Optimization Algorithm Design [实用算法设计]

This lecture was held at the Sūzhōu Institute for Advanced Studies [苏州研究院] of the School of Software Engineering [中国科学技术大学软件学院] of the University of Science and Technology of China (USTC) [中国科学技术大学] located in Sūzhōu [苏州], Jiāngsū [江苏], China [中国] in the winter semester 2010/2011 in English language.

Title:Practical Optimization Algorithm Design [实用算法设计]
Location:Sūzhōu Institute for Advanced Studies of the School of Software Engineering of the USTC; Míngdé Building, Room 324 [苏州研究院明德楼324教室]
Time:Wednesday [星期三], 14:00-17:10 (week 2 to 5 and week 7 to 17)
Lecturer:Thomas Weise [汤卫思, Tāng Wèisī], tweise@gmx.de
Assistant:Dài Dòng [代栋], daidong@mail.ustc.edu.cn, http://home.ustc.edu.cn/~daidong
Contact:Guō Yàn [郭燕], guoyan@ustc.edu.cn
Lab Class:
WeekDayTimeRoom
Week 14 to 18
Starting 2010-11-28
Sunday Evening19:00 to 22:00Room 308
Week 15 to 16Saturday Evening19:00 to 22:00Room 308
Week 18Monday Evening19:00 to 22:00Room 308
Exam:2011-01-05 between 2:00pm and 4:00pm (14:00 to 16:00)
Project:Deadline: 2011-01-05 11pm (23:00)

Material

  • Tentative Syllabus: pdf (342 kiB)
  • Teaching Material (New Book): pdf (19 MiB, version: 2011-09-19)
  • Sources: zip (4 MiB, version: 2011-01-22) [latest CVS build]

Lessons

0.Organization
  • slides (5 MiB, version: 2010-09-13)
1.Introduction
2.The Structure of Optimization
  • slides (7 MiB, version: 2010-09-16)
3.Hill Climbing and Random Walks
  • slides (3 MiB, version: 2010-09-28)
4.Difficulties in Optimization
  • slides (5 MiB, version: 2010-10-18)
5.Simulated Annealing
  • slides (5 MiB, version: 2010-10-18)
6.Genetic Algorithm
  • slides (13 MiB, version: 2010-11-03)
7.Evolution Strategy
  • slides (6 MiB, version: 2010-11-23)
8.Random Keys Encoding
  • slides (3 MiB, version: 2010-11-23)
9.Differential Evolution
  • slides (2 MiB, version: 2010-11-23)
10.Genetic Programming
  • slides (18 MiB, version: 2010-11-30)
11.Estimation of Distribution Algorithms
12.Multi-Objective Optimization and MOEAs
  • slides (15 MiB, version: 2010-12-30)
13.Swarm Intelligence
  • slides (2 MiB, version: 2010-12-28)
A.Probability Theory
  • slides (10 MiB, version: 2010-09-14)
B.Project Work
  • slides (4 MiB, version: 2010-11-09)
C.Alternative Project Work
  • slides (3 MiB, version: 2010-12-09)
D.Exam Preparation
  • slides (2 MiB, version: 2011-01-04)

Exercises and Homework

  • Homework will be given approximately every week.
  • The homework will make up 25% of the final score in the course.
  • The homework must be sent to my teaching assistant Dài Dòng [代栋] via the following email address: daidong@mail.ustc.edu.cn
  • The subject of the email must be "[xx] Student ID" where "xx" is to be replaced with the two-digit exercise number and "Student ID" with the sender's student ID.
  • Homework handed in after the deadline will be ignored and can score no points.
  • Acceptable attachments for emails are: .pdf, .txt, .doc, .jpg, .java, .c, .cpp and makefiles as well as .zip archives containing files of these types.
  • Emails containing attachments of types which are not listed, especially executable binaries, may not be processed or even be deleted.
  • Java sources must be compilable with Eclipse and C or C++ files must be compilable with GCC and accompanied by a makefile.
  • Questions regarding the homework can be sent to me via email or — in urgent cases — via QQ.
  • The homework tasks are in the electronic book. At the end of each chapter, there is a section "Tasks". This section contains the tasks to the chapter. The tasks are numbered consecutively.
  • Make sure that you always have the latest version of the book.
  • Starting with the second exercise, only solutions written in English will be accepted.
E1.Introduction
Tasks:2, 3, 12, 13, 14, 22
Deadline:2010-09-28 23:55 CST
E2.Hill Climbing
Tasks:28, part (a) of task 64 (which is worth 20 points, you do not need to do the other parts), 65
Deadline:2010-10-15 23:55 CST
E3.Random Walks and Hill Climbing
Tasks:33 (you do not need to consider adaptive walks), 34, 40, 66
Deadline:2010-10-19 23:55 CST
E4.Difficulties in Optimization
Tasks:44, 48, 50, 51, 52, 59, 60
Deadline:2010-10-26 23:55 CST
E5.Simulated Annealing
Tasks:72 OR 68 (50 points in total for both Hill Climbing and Simulated Annealing applied to the problem together)
Deadline:2010-11-02 23:55 CST
E6.Genetic Algorithm
Tasks:(79 AND 80 (not 78)) OR (81 AND 82)
Deadline:2010-11-19 23:55 CST
E7.Evolution Strategies
Tasks:85, 86
Deadline:2010-11-23 23:55 CST
E8.Random Key Encoding
Tasks:90
Deadline:2010-11-30 23:55 CST
E9.Genetic Programming
Tasks:91 OR 92
Deadline:2010-12-07 23:55 CST
E10.Estimation of Distribution Algorithms
Tasks:Any combination of Tasks 99 to 103 that adds up to 50 points.
Deadline:2010-12-14 23:55 CST