Tuning a Simulated Annealing Heuristic for Irregular Polyomino Tiling

This paper describes using simulated annealing and integer programming to the problem of irregular polyomino tiling. Simulated annealing (SA) has many parameters which significantly influence on its effectivity. Thus, their values need to be carefully selected when applied to a specific problem. The most obvious way to tune the SA parameters is a full factorial experiment which can consume a lot of time. In this paper we suggest using the Taguchi method in order to tune SA parameters to solve a problem of irregular polyomino tiling. This approach allows to find the most influential of the parameters under consideration. One well-known application of the irregular polyomino tiling problem is the phased array antennas design. The computational results and antenna simulations show the efficiency of proposed SA tuning method.