Solution to 2D Rectangular Strip Packing Problems Based on ACOs

As a category of stock cutting problems, 2D rectangular strip packing problems are NP-hard in the strong sense, which are mainly resolved by simple heuristic algorithms and intelligence optimum algorithms. Among those methods, many researchers approached 2DRSPP by genetic algorithms and simulated annealing. This paper, using max-min ant system, generates the input sequence of packing following by the improved placement strategy resulting in the phenotype of solution. Meanwhile a set of rules are added while selecting the next rectangular from the candidate list for speeding the constringency. The selection of initial rectangular is in proportion with results combined with random. Placement strategy is divided into two stages. The first is based on level-orientation, and the second uses sliding methods. With two benchmarks on these approaches, the final test shows that this algorithm is an efficient one on results and time.