GPLAB is a Genetic Programming toolbox for MATLAB. Most of its functions are used as "plug and play" devices, making it a versatile and easily extendable tool, as long as you have minimum knowledge of the MATLAB programming environment. Here is an illustration of how it works. For details, please read the user's manual. Note: the operational structure is now somewhat outdated, but it may still be useful anyway. The features and references are being updated (2015).

Some of the features of GPLAB (good or bad):

• 3 modes of tree initialization (Full, Grow, Ramped Half-and-Half) + 3 variations on these [2]
• several pre-made functions and terminals for building trees
• dynamic limits on tree depth [1] or size [2] (optional)
• resource-limited GP (variable size populations) [3] [4] (optional)
• dynamic populations (variable size populations) [5] [6] [7] [8] (optional)
• 4 genetic operators (crossover, mutation, swap mutation, shrink mutation)
• configurable automatic adaptation of operator probabilities [11] (optional)
• steady-state + generational + batch modes, with fuzzy frontiers between them
• 5 sampling methods (Roulette, SUS, Tournament, Lexicographic Parsimony Pressure Tournament [9], Double Tournament [10])
• 3 modes of calculating the expected number of offspring (absolute + 2 ranking methods)
• 2 methods for reading input files and for calculating fitness (symbolic regression and parity problems + artificial ant problems)
• runtime cross-validation of the best individual of the run (optional)
• offline cross-validation or prediction of results by any individual (optional)
• 4 levels of elitism
• configurable stop conditions
• saving of results to files (5 frequency modes, optional)
• 3 modes of runtime textual output
• runtime graphical output (4 plots, optional)
• offline graphical output (5 functions, optional)
• runtime measurement of population diversity (2 measures, optional)
• runtime measurement of average tree level, number of nodes, number of introns, tree fill rate (optional)
• 4 demonstration functions (symbolic regression, parity, artificial ant, multiplexer)

GPLAB does not implement:

• multiple subpopulations
• automatically-defined functions

References:

[1] Silva S, Almeida J (2003). Dynamic Maximum Tree Depth - A Simple Technique for Avoiding Bloat in Tree-Based GP. In Cantú-Paz E, et al. (eds), Proceedings of GECCO-2003, 1776--1787.

[2] Silva S, Costa E (2004). Dynamic Limits for Bloat Control - Variations on Size and Depth. In Deb K, et al. (eds), Proceedings of GECCO-2004, 666--677.

[3] Silva S, Silva PJN, Costa E (2005). Resource-Limited Genetic Programming: Replacing Tree Depth Limits. In Ribeiro B, et al. (eds), Proceedings of ICANNGA-2005, 243--246.

[4] Silva S, Costa E (2005). Resource-Limited Genetic Programming: The Dynamic Approach. In Beyer H-G, et al. (eds), Proceedings of GECCO-2005, 1673--1680.

[5] Tomassini M, Vanneschi L, Cuendet J, Fernandez F (2004). A New Technique for Dynamic Size Populations In Genetic Programming. In Proceedings of CEC-2004, 486--493.

[6] Cuendet J (2004). Populations dynamiques en programmation génétique. Université de Lausanne, Université de Genève.

[7] Rochat D (2004). Programmation Génétique Parallèle: Opérateurs Génétiques Variés et Populations Dynamiques. Université de Lausanne, Université de Genève.

[8] Rochat D, Tomassini M, Vanneschi L (2005). Dynamic Size Populations in Distributed Genetic Pogramming. In Keijzer M, et al. (eds), Proceedings of EuroGP-2005, 50--61.

[9] Luke S, Panait L (2002). Lexicographic Parsimony Pressure. In Langdon WB, Cantú-Paz E, Mathias K, et al. (eds), Proceedings of GECCO-2002, 829--836.

[10] Luke S, Panait L (2006). A comparison of bloat control methods for genetic programming. Evolutionary Computation 14(3): 309--344.

[11] Davis L (1989). Adapting operator probabilities in genetic algorithms. In Schaffer JD (ed.), Proceedings of the Third International Conference on Genetic Algorithms, 61--69.