Adaptive iterative improvement GP-based methodology for HW/SW co-synthesis of embedded systems

The paper presents a novel adaptive genetic programming based iterative improvement algorithm for hardware/software co-synthesis of distributed embedded systems. The algorithm builds solutions by starting from suboptimal architecture (the fastest) and using system-building options improves the system’s quality. Most known genetic programming algorithms for co-synthesis of embedded systems are built choosing fixed probability. In our approach we decided to change the probability during the work of the program