A Crossover Operators in a Genetic Algorithm for Maritime Cargo Delivery Optimization

Abstract

Maritime cargo delivery accounts for over 80% of the world's trade and contributes about 3% of the world's gross domestic product. Here, we focus on the problem of minimizing the maritime cargo delivery cost by an amount equivalent to the sum of the tour lengths of feeders used for the delivery. We formulate maritime cargo delivery cost reduction as a multiple traveling salesman problem and use a genetic algorithm to solve it. In addition to minimizing the route length, the algorithm indirectly reduces the number of feeders. To increase the performance of the genetic algorithm, we implement a 3-point crossover operator, which takes three chromosomes and returns slightly more complex crossover mutations than the known 2-point crossover operator. These two operators must be used in confluence. We propose to run both the 2-point crossover algorithm and the 2-point-and-3-point crossover algorithm in parallel and select the route with the shortest length. The route length is cut down by a few percent, which makes a big difference in how much it costs to ship cargo by sea.

Keywords

• maritime cargo delivery
• tour length
• genetic algorithm
• crossover