<![CDATA[There are several ways to solve sudoku puzzles with a computer that is Trial and Error, the adaptation of human intuition and using optimization algorithms. Particle Swarm Optimization (PSO) is one of the methods of optimization are found in a group of intelligence, which is based on the population of the algorithm searches and begins with a random population called particle. This study was conducted to test the PSO method to solve sudoku puzzle with several levels of dimensions 4x4 up to 25x25. Testing is done by changing the number of particles and see its effect on the value iteration and waiting time. The results of this study indicate several things. First, the greater the value of the particle yield tends to decrease the number of iterations. Second, the same number of iterations is not necessarily produce the same waiting time. Third, that the smaller the number of iterations is not necessarily produce a longer waiting time. Fourth, the larger size is not necessarily produce a longer waiting time. Fifth, the puzzle that has many cells with the possibility of a valid number or bit combinations are possible, PSO will shorten the number of iterations, but for the puzzle that has many possible numbers of cells with more than one or many combinations are possible, PSO not a lot of influence. Sixth, the large dimensions there are significant differences in waiting times between the number of particles 1 with the number of particles 2 to 10. Seventh, the dimensions of the primes tend to have a greater number of iterations when compared with other dimensions. The value of velocity is used to determine the value of randomization number of particles, so that the particles are sometimes not always move forward or better value.]]>
