Simultaneous Optimistic Optimization

SOO (Simultaneous Optimistic Optimization)

Author: Rémi Munos

This is a blackbox function optimization toolkit that finds the global optimum of a function given a finite budget of deterministic evaluations. The algorithm does not require the knowledge of the function’s local smoothness around its global optimum.

Reference: Rémi Munos: Optimistic optimization of a deterministic function without the knowledge of its smoothness, in NIPS 2011

• Paper: http://chercheurs.lille.inria.fr/~munos/papers/files/opti2_nips2011.pdf
  * Software: SOO (in C) v1.0
  * Demo videos:
  • http://www.youtube.com/watch?v=mUgNtRUDGV4
  • http://www.youtube.com/watch?v=b_AkkfkygLU
  • http://www.youtube.com/watch?v=J9Ei1JkxBw4
  • http://www.youtube.com/watch?v=5g2ltjJoB40

Getting started

• Unzip the archive into a directory of your choice.
• go to soo directory and type ‘make’ to compile all files (requires libraries draw, utils and extra)
• Example: run ‘./soo 100″ (100 is the number of function evaluations) or alternatively go to ‘LINUX/Debug’ and run one of the soo_name_of_function program
• Several functions f(.) are commented out in main.c so you can choose another function by uncommenting it or by writing your own f(.) function