A set of methods for enhancing the efficiency of information processing in intelligent decision support systems

This section of the study proposes a set of methods to enhance the efficiency of information processing in intelligent decision support systems.

The authors suggest the following methods:
– a method for managing information flows in intelligent decision support systems using a population-based algorithm;
– a method for evaluating the timeliness of processing diverse data types in decision support systems;
– a method for assessment and forecasting in intelligent decision support systems.

The novelty of the proposed methods lies in:
– determining the initial population of agents and their starting positions in the search space by considering the degree of uncertainty in the initial data about information flows within intelligent decision support systems;
– accounting for the initial velocity of each agent, which enables prioritization of search tasks within the respective search space;
– universality of agent feeding search strategies, allowing the classification of conditions and factors that influence the management of information flows in intelligent decision support systems;
– ability to explore solution spaces described by atypical functions through the application of agent movement technique selection procedures;
– capability to search for solutions simultaneously in multiple directions;
– potential for deep learning of agents' knowledge bases;
– ability to calculate the required amount of computational resources to be engaged in cases where existing resources are insufficient for necessary computations;
– consideration of the type of uncertainty in the data circulating within decision support systems;
– implementation of adaptive strategies for solution space searches by the population agents;
– prioritization of search tasks by population agents;
– initial placement of population members considering the type of uncertainty;
– application as a universal tool for analyzing the timeliness of processing diverse data types in decision support systems;
– verification of the adequacy of the obtained results;
– avoidance of the local extremum problem;
– use of a new type of fuzzy cognitive temporal models focused on multidimensional analysis and forecasting of object states under conditions of uncertainty;
– ability to combine elements of artificial neural networks;
– capability to train individual elements of artificial neural networks;
– data computation within a single epoch without the need to store previous calculations;
– prevention of error accumulation during the training of artificial neural networks as a result of processing incoming information.