Intelligent model distribution method, distribution system and application system

Abstract

An intelligent model distribution method, distribution system and application system relate to the field of optical communication, and the method comprises the steps of determining main features, feature types and feature attributes of an intelligent model in each distribution scheme, constructing a feature table in combination with feature values obtained by multiple tests, establishing an evaluation matrix, and calculating a normalized foreground matrix of the distribution scheme; calculating an objective weight under the condition of maximizing the allocation value, calculating an evaluation target according to the subjective weight and the objective weight, gradually updating the objective weight by adopting a gradient method until the evaluation target is smaller than a set value, andcalculating the allocation value of the corresponding allocation scheme by taking the objective weight as an equilibrium weight at the moment; and comparing the distribution values of different distribution schemes, and selecting the distribution scheme with the maximum distribution value. According to the invention, the optimal scheduling of the intelligent model can be obtained, and the efficiency maximization is achieved.

Description

technical field [0001] The invention relates to the field of optical communication, in particular to an intelligent model distribution method, distribution system and application system. Background technique [0002] In the existing research, the expected utility theory is mainly used to study the multi-attribute decision-making problem. But expected utility theory has an inherent flaw: it assumes that all decision makers are perfectly rational and always seek to maximize utility. Therefore, the conclusions obtained from it are usually inconsistent with the actual decision-making behavior of decision-makers under uncertainty, such as the famous Allais paradox and Ellsberg paradox. The prospect theory uses the value function v(x) and the decision weight function w(p) to describe the individual's choice behavior. The value function is related to the selection of the reference point, and the prospect value is jointly determined by the value function and the decision weight fu...

AI Technical Summary

Problems solved by technology

But expected utility theory has an inherent flaw: it assumes that all decision makers are perfectly rational and always seek to maximize utility

The disadvantage of this method is: the use of fuzzy recognition theory is aimed at a single characteristic attribute value, and the mixed multi-attribute situation encountered in the field of optical communication intelligent model scheduling

[0004] To build a model system for allocation of equipment and management and control platforms, an alternate iterative algorithm based on Lagrangian dual decomposition and linear programming is currently used. The disadvantage of this method is that it ignores that the actual gain and loss will change due to different actual situations.

The disadvantage of the above method is: the weights are all derived by the decision maker (or analyst) at one time, lacking dynamic programming

There are few researches on intelligent model scheduling in the field of optical communication, and the public resource scheduling methods also have defects in varying degrees. It is difficult to achieve ideal results when directly applied to optical communication intelligent model scheduling, and the efficiency is low.

Images