In one word

The best place to learn about the features and approaches of EO with the help of examples is to look at the tutorial.

Once you have understand the Overall Design of EO, you can search for advanced features by browsing the modules page.

Introduction

EO is a template-based, ANSI-C++ evolutionary computation library which helps you to write your own stochastic optimization algorithms insanely fast.

It contains classes for almost any kind of evolutionary computation you might come up to - at least for the ones we could think of. It is component-based, so that if you don't find the class you need in it, it is very easy to subclass existing abstract or concrete classes.

Designing an algorithm with EO consists in choosing what components you want to use for your specific needs, just as building a structure with Lego blocks.

If you have a classical problem for which available code exists (for example if you have a black-box problem with real-valued variables), you will just choose components to form an algorithm and connect it to your fitness function (which computes the quality of a given solution).

If your problem is a bit more exotic, you will have to code a class that represents how your individuals (a solution to your problem) are represented, and perhaps some variations operators, but most of the other operators (selection, replacement, stopping criteria, command-line interface, etc.) are already available in EO.

Overall Design

EO is a framework. It is oriented toward facilitating the design of adhoc evolutionary algorithms. It is not (at the moment) a complete library of algorithms ready to use on canonical problems.

If you have a well-known problem and want to solve it as soon as possible, try another software. If you have a real problem and want to build the best evolutionary algorithm to solve it, you've made the good choice.

Bascially, EO manipulate "individuals" with a "fitness", that is objects encoding a solution to a given optimization problem, associated with the quality of this solution. The fitness is defined in the EO class, but the representation of a solution cannot be as generic. Thus, EO massively use templates, so that you will not be limited by interfaces when using your own representation.

Once you have a representation, you will build your own evolutionary algorithm by assembling Evolutionary Operators in Algorithms. In EO, most of the objects are functors, that is classes with an operator(), that you can call just as if they were classical functions. For example, an algorithm is a functor, that manipulate a population of individuals, it will be implemented as a functor, with a member like: operator()(eoPop<EOT>). Once called on a given population, it will search for the optimum of a given problem.

Generally, operators are instanciated once and then binded in an algorithm by reference. Thus, you can easily build your own algorithm by trying several combination of operators.

For a more detailled introduction to the design of EO you can look at the slides from a talk at EA 2001 or at the corresponding article in Lecture Notes In Computer Science, 2310, Selected Papers from the 5th European Conference on Artificial Evolution: