Genetically designed heuristics for the bin packing problem

The bin packing problem (BPP) is a real-world problem that arises in different industrial applications related to minimization of space or time. The aim of this research is to automatically design an algorithm that places a collection of objects into the minimum number of fixed-size bins. For generating this heuristic we use Genetic Programming (GP) with a special set of terminals. The evolved strategy is compared to Best Fit Descending (BFD), a well-known heuristic for the bin packing problem. The results emphasize that evolved GP heuristics can perform equally and sometimes even better than BFD for the considered test problems.