零件编码码位定量计估中的信息论方法

Quantitative Evaluation of Part Classification Code-Bit Based on Information Theory

零件分类编码中,编码码位的信息含量与码位各状态出现的概率分布有内在联系。本文以Shan- non信息论为基础,从编码信息含量的角度出发,提出码位信息剩余度、码位不均衡损失函数等概念,建立了编码码位各状态调整或合并的准则,给出优先级别为信息量、码值距离的状态合并的一种算法,从而使码位满足信息量与均衡方面的要求,为减少对象冗余提供依据;并举例说明该方法在紧闹件编码中的应用。

In part classification coding system, the information embodied in codes has internal relation with the probability distribution of the code value. Based on C. E. Shannon＇s information theory, the conception of residual en- tropy of bit, non-equilibrum loss function of bit, some rules of merger and adjustment of bit value were proposed, An algorithm of bit value merger was proposed based on priority level of information and bit-distance so that bits can satisfy the requirement of both information amount and equilibrium, and facilitate the reduction of parts＇redundancy. Taking fastener＇s codes as an examole the appkication of the method is illustrated.