A METHOD FOR SOLVING THE TRANSPORTATION PROBLEM FOR SEAPORTS BASED ON THE USE OF INTELLIGENT AGENTS

Abstract

The purpose of the article is to develop a method for solving a transportation problem for seaports based on the use of intelligent agents, which allows obtaining solutions that are as close to optimal as possible. This goal is achieved by formulating a generalized model routing problem, developing procedures for forming coalitions of agents and distributing the desired actions of agents with the possibility of changing success ratings. It has been established that in modern times, an effective way is to develop intelligent software systems based on multiagent systems and rules used in maritime transport technologies. In general, agent-based modelling is widely used to study complex dynamic systems. An intelligent agent, which has the properties of autonomy, social behaviour, reactivity, and proactivity, is such a software object. In this paper, the term "intelligent transport agent" is used to refer to an agent that solves routing problems, as well as can move and characteristics of speed and carrying capacity. The task of improving the quality of solving high-dimensional routing problems based on the use of coalitions of intelligent transport agents, with assessments of the success of actions that change during activity, is relevant within the framework of maritime transport technologies. The most significant result is a method for solving the transport problem for seaports based on the use of intelligent agents, which involves the formation of coalitions of agents and the distribution of desired actions with the possibility of changing success ratings. Within the framework of the corresponding method, procedures are proposed to ensure high quality of solving the transportation problem. At the same time, the article establishes the computational complexity of the algorithms used to implement these procedures. The presented approach to creating agents allows for increased flexibility in their behaviour, easy modification, and separation of the program code from the description of actions that determine behaviour. An approach based on improving the characteristics of agents because of the actions performed by agents can be used to bring the obtained solutions closer to optimal results.