Design and Implementation of Simulated Annealing Optimization Algorithm Based on TOC

When the traditional simulated annealing algorithm solves the traveling salesman problem (TSP), with an increase in the size of the problem, the time required increases and the quality of the solution decreases. To this end, this paper takes advantage of the parallelism and bit-wise assignability of ternary optical computers and proposes a parallelization strategy that divides an optical processor into many small optical processors and then distributes the initial solution to the problem to each small optical processor. On the optical processor, the optimal solution is independently searched for. Due to the parallel search of many processors, the convergence speed of the algorithm and the global optimization ability are effectively improved. Experimental tests show that the parallel strategy can achieve a better convergence effect and effectively improve the quality of understanding.