Genetic algorithms - synthesis of finite state machines

Genetic algorithms (GAs) are a stochastic, non-derivative optimization method. They use populations of acceptable solutions (genes) of the given problem, which evolve toward optimum. The paper introduces GAs as a method for the synthesis of the activation function of flip-flops in finite state machines. The genes in standard GAs are Boolean vectors. When JK and RS flip-flops are used in the synthesis of finite state machines, there are undefined variables in the activation signals. When the finite state machine is of high order, the Quine-McClusky method is used, which requires exact values of the variables. At this stage, the GAs are used to find the optimal set of variables, in terms of simplifying the description.