Genetic Algorithm to generate Binary Code

Hello,

I am using GA to generate an optimal binary shutter sequence for a coded exposure camera. Since I want to generate a code of 0 and 1, I used IntCon to force all the components of my vector to be integers and set lower bound 0 and upper bound 1 for the elements. However, using IntCon restricts the range of functions I can use for Mutation, Selection, Creation and so on. At the moment, I am trying to reproduce the method the authors use in this paper: "Optimal binary sequences for coded exposure photography using genetic algorithm" K.Huang, J.Hou, J.Zhang, published in IEEE 2014

http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=6986200&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D6986200

They say, quote, "we used the stochastic uniform selection function, scatted crossover function and adapt feasible mutation function in GA." But some of these are overridden by ga in integer problems. Moreover, the ga automatically uses a penalty function instead of the fitness function...

Is there any way I can restrict the components of my vector to just 0 and 1 without using IntCon, such that I can use the same methods as the paper above?

Many thanks,

Stefania

Answer 1

[Walter Roberson]

I cannot seem to find it now, but I recently saw someone answer a related question. The solution in that case was to not specify integer constraints, because doing that restricts what you can do. Instead, the way to proceed was to provide your own custom functions for initial population and mutation and crossover which always return integer results according to your needs.

[Stefania]

Thanks for the suggestion!