The Pacific Ocean route optimization by Pittsburgh-style Learning Classifier System

This paper proposes the optimization method which extends Pittsburg-style Learning Classifier System (LCS) for Pacific Ocean route. In detail, the following extensions are introduced: (1) the unrealistic route deletion, (2) the route integration, and (3) the route rest time minimization and the anchor order change. To investigate the effectiveness of the proposed methods, this paper applies them into LCS to optimize the Pacific Ocean liner route using the actual transportation data. The intensive simulations have revealed following indications: (1) the generated routes using the proposed methods can produce the feasible routes that are had to be found by the conventional method; and (2) our proposed methods contribute to creating the effective route set which has the short rest time, a small number of vessels, and high profit.