Weighted random mixed semantic method for combat effectiveness evaluation

Abstract

The invention discloses a weighted random mixed semantic method for combat effectiveness evaluation. The method comprises the following steps of evaluation index system construction and index empowerment, evaluation index quantification and remapping, weighted random mixed semantic model establishment and data substitution inference. According to the method, under the framework of the Bayesian network, a weighted random mixed semantics is used for describing a combat effectiveness evaluation process, a quantitative index and a qualitative index can be mixed and incorporated, the quantitative index and the qualitative index only have differences in a random semantic level, and potential information provided by the qualitative index can be fully utilized; meanwhile, the method is substantially a Bayesian posterior inference method, so that the method has the advantage that a Bayesian method allows missing data, quantifiable indexes are reflected as observation value nodes in a model, missing indexes are reflected as unknown parameter nodes, and the parameter nodes are not used as evidence of model inference, so that the model inference efficiency is greatly improved. Therefore, the missing index has no influence on the evaluation process and the evaluation result.

Description

technical field [0001] The invention relates to the technical field of combat effectiveness evaluation, in particular to a weighted random mixed semantic method for combat effectiveness evaluation. Background technique [0002] Combat effectiveness measures the extent to which a specific combat operation or combat entity achieves a predetermined combat goal under certain constraints. Combat effectiveness evaluation is a quantitative estimation process for this level of abstraction. It constructs an evaluation index system and evaluation method in a targeted manner. It can quantify, synthesize and rank key operational elements and operational processes. Therefore, combat effectiveness evaluation has great guiding significance for practical issues such as combat plan formulation, combat plan optimization, and combat system construction. [0003] At present, the Analytic Hierarchy Process, the Fuzzy Comprehensive Evaluation Method and their various variants have been widely us...

AI Technical Summary

Problems solved by technology

[0004] However, with the continuous development of weaponry and combat styles, the construction environment of combat effectiveness evaluation (such as evaluation objects, evaluation criteria, data collection methods, etc.) is becoming more and more complex and diverse, and these traditional methods have gradually revealed certain limitations.

First, the inability to handle qualitative indicators

Qualitative indicators are sortable but non-quantitative indicators with discrete label values ​​such as "good, bad", "high, low", "strong, weak", etc., and these methods require that the indicator values ​​must be quantifiable real Values, qualitative indicators are either ignored or forced to assign a specific value, which is obviously inappropriate; secondly, missing indicators cannot be dealt with

There are many factors that cause missing indicators, such as acquisition sensor failure, data transmission link failure, storage medium damage, low data quality, etc. The external performance is an indicator with a value of "null".

These methods generally adopt an empirical filling strategy when dealing with missing indicators, that is, filling the minimum value, maximum value or average value of historical data, etc., which is obviously imprecise

Images