Ship Route Designing for Collision Avoidance Based on Bayesian Genetic Algorithm

Maritime accidents frequently occurred because of human errors. It is important to generate safe route automatically in collision situations. Genetic algorithm is a new method for this purpose. However, the International Regulations for Preventing Collision at Sea have not been taken into account in previous researches. In this paper, Bayesian model is employed to calculate the post probability of each trajectory according to the rules and general practice of seaman. The fitness function is formed with path cost and information content. Experiments illustrate that evasion routes generated by Bayesian genetic algorithm are more rational than that of pure GA, considering the rules for preventing collision and general practice of seaman.