January 27, 2020
Grammatical Evolution, Evolutionary Computing
Prof. Conor Ryan
Dara O Shayda (LoG)
Imran Gholizadeh (CCN)
Grammatical Evolution has been a quantum leap in human cognition of the evolutionary systems which its mutations (Crossovers) output valid words in some specific grammar. The fact that such systems could be programmed and used to actually solve problems is quiet an achievement. However the underlying operators and the grammars have been quite simple and to extend the grammar one needs to handcraft quite a large new grammar. Funge is a Functional based Grammatical Evolution which uses all the functional facilities of Mathematica and access to its vast mathematical functions to create tiny grammars with amazing expansions of operators and functionals. Moreover there are no limitations to the input array i.e. being integers, now they could be any data type e.g. images or audio output-ing words valid in some grammar! Variables and Expressions and Data and geometries and formulas and images and audio and many other data types could be confused with each other to produce a powerful programming environment for Funge to generate code. Random Operators e.g. Random Differential Operators and quite sophisticated random structures could be created with varying statistical distributions. Finally grammar itself is also a system variable and could be evolved as do the other parts of the Funge.