Binary Fick's Law Algorithm based on a new repair optimization operator for solving hardware/software partitioning

To efficiently solve the hardware/software partitioning problem (HW/SW) using Fick's Law Algorithm (FLA), this paper first proposes a Binary Fick's Law Algorithm (BFLA) based on the Sigmoid function. To further improve the quality of HW/SW solutions obtained by BFLA, a new Improved Greedy Repair and Optimization Operator (IGROM) is developed based on the greedy strategy of "hardware cost/(software cost + real-time communication cost)". Comparative experiments on 8 standard HW/SW instances verify that BFLA outperforms existing methods in terms of optimal solution quality, average performance, and stability, making it the most effective algorithm for solving HW/SW problems currently.