Optimization of the distribution of limited water resources by methods of evolutionary genetic programming

The purpose of this work is to improve the system of water distribution for irrigation in conditions of water scarcity on the basis of multi-criteria optimization by the method of evolutionary genetic programming. The methodological basis of the research was the information -analytical approach; optimization modeling; artificial intelligence methods (genetic algorithms of evolutionary programming). A universal multi-criteria, nonlinear optimization objective function has been developed and tested, including criteria for the effectiveness of the operational organization and water user farms. These include: the maximum area of irrigated land, the income of the water management organization and the gross volume (in monetary terms) of crop production from irrigated land. The scientific and methodological directions of the use of genetic algorithms in the optimization of the distribution of limited water resources on inter-farm irrigation systems are considered. The possibility and effectiveness of using a genetic algorithm to optimize a multi-criteria objective function is established. Choosing a solution based on the indicators of multiple imitations of competition and improvements increases the scale of a complex solution search space and, as a result, universality allows you to get results when traditional methods do not work or their implementation requires an unacceptably long time. Testing of the developed model, carried out on the data of the Gorodishchenskaya irrigation system of the Volgograd region, revealed a projected increase in the effectiveness of the management decision by 10% in comparison with the traditionally practiced approach. The result is achieved by improving the quality of planning and management of water distribution in conditions of water scarcity on the basis of information support of decisions by innovative methods of management theory: multi-criteria modeling using genetic programming based on artificial intelligence methods.