Abstract

CodeMonkey-GA (CM) is a GUI driven software development platform that enables non-experts and experts alike to turn an evolutionary algorithm design into a working Java program, with a minimum amount of manual coding. CM is provided as a framework and plug-in application for the Eclipse platform for non-commercial uses.
We compare some of the most popular frameworks and platforms for evolutionary computation. We discuss their shortfalls and justify the need for still another platform. Hence, we present CodeMonkey-GA: its concept, internal architecture and design. We provide an overview of the graphical user interface (GUI) of the platform followed by examples of evolutionary algorithm applications, all generated using CodeMonkey’s Eclipse application.
Through several examples we demonstrate the ease of use and (to some degree) the applicability of the CM application. The Ackley function is a well-known test function for optimization; the Traveling Salesman Problem is a famous example of NP-Complete problems; the Knapsack problem is an example of combinatorial optimization. In all three cases, CM is used to develop working Java programs that provided satisfactory solutions, which are as good as, or better than the given solutions. Critically, in all cases, not a line of code was entered or altered – bar the fitness function – by the user.