The Berth Allocation Problem (PAB) consists of ships programming and allocation to the mooring areas along a quay. PAB is modeled as a Multi-Depot Vehicle Routing Problem with Time Windows. A recently proposed method denominated ATP/PL is used to solve it. This method consists of the application of the Population Training Algorithm (ATP) with Linear Programming (PL) for Column Generation. These methods are applied in an interactive way, where ATP uses PL relaxation information for the generation of good columns, and PL is used for solving a Set Partitioning Problem, with an additional restriction (PPC+), formed by these columns. The computational results are obtained through instances used in a recent work and compared against results from CPLEX and other method found in literature.
berth allocation problem; population training algorithm; column generation
