Skip to main content Skip to local navigation

Experimental analysis of attraction basins in multi-modal search spaces

Experimental analysis of attraction basins in multi-modal search spaces

Home » Information Technology » ITEC 4000: Independent Research Project » Experimental analysis of attraction basins in multi-modal search spaces

Experimental analysis of attraction basins in multi-modal search spaces

Project Supervisor: Stephen Chen
Desirable Background: Strong mathematical, programming, and analytical skills (indicated by A grades in Math and Programming courses)
Required Background: An interest in pursuing a graduate degree
Project Session: Fall, Winter, Summer
Project Description:

The typical conic shape of attraction basins in multi-modal search spaces means that the linear scale of fitness evaluations is mapped against an exponential volume of actual solutions in comparison -- the range of the top 25% in possible fitness for solutions likely translates to less than 1% of the actual volume of the attraction basin.  The effect of this feature will be studied using Monte Carlo simulations on benchmark problems (in one to thirty dimensions) and through the application of benchmark metaheuristics (mostly likely Particle Swarm Optimization) to these problems.

 

Categories: