Development of a flexible hardware core for genetic algorithm

Abstract

A hardware design for genetic algorithm (GA) can implement only one specific cost function of a problem at a time. Actually, different GA applications require different GA hardware architecture. The development of a flexible very-large-scale integration (VLSI) for GA has been proposed in this paper. For the hardware architecture, we has develop on a random number generator (RNG), crossover, and mutation based on flexibility structure. This structure can dynamically perform to the 3 types chromosome encoding: binary encoding, real-value encoding, and integer encoding. The overall structures has been designed and synthesized by VHDL (VHSIC hardware description language), simulation by ModelSim program, and then implemented on FPGAs (Field programmable gate arrays). This hardware architecture that our design work very well flexible for the 3 groups problem examples: combinatorial optimization problems, function optimal problems, and part planning optimization problems.