Abstract
This paper aims to ascertain the optimum values for two fitness function parameters within a process mining genetic algorithm; the o parameter, which reduces the likelihood of process models with extra behaviour being selected and the parameter, which restricts the selection of models containing duplicate tasks. The experiments conducted in this research also include the use of a decaying rate for the mutation operator in order to promote greater accuracy in the mined process models. The paper concludes that the optimum setting of the fitness function parameters will in fact vary depending on the constructs found in each process model. This paper finds that a higher value for one of the fitness function parameters allows for simple process constructs to be mined with greater accuracy. The use of a decaying rate of mutation is also found to be beneficial in the correct mining of simple processe.