Gene expression programming for induction of finite transducer

This paper presents an alternative method for solving the problem of finite transducers using gene expression programming (GEP). Each individual in the GEP system represents a Mealy machine with outputs for each state. Be means of roulette-wheel sampling, individuals are chosen for the next generation and are put through a series of genetic operators which seek to change the mark-up of the individual to better fit the selection environment/ fitness sets. The system was tested with five problems to show its effectiveness and success at solving all of those problems.