Stochastic Optimization

• Home
• Github repository

Syllabus

This class covers stochastic methods of optimization, primarily simulated annealing, evolutionary strategies, and genetic algorithms. The class is 10 hours total and uses HTML presentations and Jupyter notebooks in Python for exercises. The evaluation for this class is based on quiz responses during the three classes.

Schedule
16/10	Introduction and simulated annealing	Continuous optimization, random search, simulated annealing
17/10	Evolutionary Strategies	Population-based methods, 1+1 ES, CMA-ES
07/11	Genetic Algorithms	Genetic Algorithm, Multi-Objective Optimization, NSGA-II

Resources

• Pymoo
• pycma
• CMAES
• DAEP
• gplearn
• evosax
• ECJ

The 2nd year elective class EISC217: Evolutionary Computation goes into further detail on many of these same topics and introduces new topics such as genetic programming and quality diversity.

The Introduction to Evolutionary Computing book by A. E. Eiben is recommended as reading for this class.