The Advanced GA Options page allows you to adjust the genetic algorithms (GA) settings for speed and efficiency. The default values provided will produce reasonable results over a wide range of modeling circumstances. The advanced options will allow you to fine tune the GA process if required.

Click on the dialog box below for additional information:

About fmGA implemented in Designer

Since Goldberg (1989) introduced GA approaches for engineers, there have been many variants of the GA technique developed thus far. The fast messy GA (fmGA) implemented here is one of the most competent types of GA delivering reliable solutions (*Goldberg et al, 1993).

The fmGA proceeds in two phases of genetic operations. It starts with an initial population of full-length strings and followings by a filtering process including building blocks and reproduction. It identifies short strings of higher fitness by randomly deleting genes from the initial strings. The building block filtering process continues until the length of strings is reduced to a prescribed order.

The identified short strings are used to form individual solutions. An individual solution is produced by “cut” and “splice” operations, which are different from the standard GA crossover and mutation operations. “Cut” divides one string into two strings, whereas “splice” concatenates two strings to form one individual solution. Both genetic operations are designed to effectively exchange the building blocks for composing the (near) optimal solution.

The fmGA combines the building blocks identification and reproduction phases into one artificial evolution process and continues over a number of outer iterations of solution initialization, filtering, and reproduction. The iterations are continued until the (near) optimal solution is identified in a population.