Packing Unequal Disks in a Circular Container Using a Population Based Algorithm

There are a lot of researches focused on the problem of unequal disks packing for its simplicity in definition and plenty of applications. Because it is NP-hard, most of them are heuristics. up to now, one of the best heuristics is iterated Tabu search. in this paper, we enhance this method by making it population based. We maintain a queue for the best k solutions found so far, and at each time, we choose a solution from it according to an elaborately set probability, perturb the solution and optimize it by adaptive Tabu search, finally, update the queue using the result of the optimization. to see the efficiency of our algorithm, we use it to solve a couple of benchmark problems, and the computational results showed its efficiency by beating several world records.