Global Minimum of a Surface

Requires a Wolfram Notebook System
Interact on desktop, mobile and cloud with the free Wolfram Player or other Wolfram Language products.
This Demonstration applies a simple evolutionary algorithm to find the global minimum for a closed and bounded surface in 3D space.
Contributed by: Daniel de Souza Carvalho (September 2015)
Open content licensed under CC BY-NC-SA
Snapshots
Details
Description
1. For the initial generation, 10 or more points are randomly selected in the plane below the surface. 2. Two pairs of points are selected randomly from the initial population. Each pair produces a new point halfway between them. 3. From those two new points, the one below the surface at a lower height is selected. 4. Steps 2 and 3 are repeated until there are enough points for the next generation.
Rastrigin surface:
sinc surface:
exp surface:
Rosenbrock surface:
References
[1] K. A. De Jong, Evolutionary Computation: A Unified Approach, Cambridge, MA: MIT Press, 2006.
[2] C. Jacob, Illustrating Evolutionary Computation with Mathematica, San Francisco: Morgan Kaufmann, 2001.
Permanent Citation