区间概率及不确定语言变量
1.区间概率
定义1:
n个实数区间[Li,Ui](i=1,2,…,n),若满足:0≤Li≤Ui≤1(i=1,2,…,n),则可以用来定义Ω中基本事件相应的概率,称之为n维区间概率,简记为n-PRI,为方便起见,引入向量:L=(L1,L2,…,Ln)T,U=(U1,U2,…,Un)T,则n-PRI又可记为n-PRI(L,U)。
定义2:
给定一个n-PRI(L,U),若存在一组正实数p1,p2,…,pn,且有
则称n-PRI(L,U)是合理的;否则n-PRI(L,U)是不合理的。
引理1:一个n-PRI(L,U)是合理的,当且仅当
如果n-PRI(L,D)是合理的,则可以进一步将概率区间[Li,Ui](i=1,2,…,n)精确化,得到概率区间为:
当然,一个不精确的n维概率区间可以通过上述公式将其精确化。
例子:
一个4维的概率区间{[0.1,0.2],[0.2,0.7],[0.3,0.4],[0.1,0.5]}是不精确的,通过上述公式将其精确化为:{[0.1,0.2],[0.2,0.5],[03,0.4],[0.1,0.4]}。
2、不确定语言变量
决策者在对事物进行定性评价时,一般需要事先指定语言评价集。设语言评价集为:
S={sα|α=0,1,…,L-1},其中,sα表示语言变量,L 为奇数。S具有如下性质:
在实用中L一般取3,5,7,9等。其表示为:
为了减少语言决策信息在运算过程中造成的丢失,把原有的离散语言评价集S=(s0,s1,…,sL-1)拓展成连续语言评价集S={sα|α∈[0,q]},其中,q是一个充分大的实数。若sα∈S,我们称sα为本原术语;否则称sα为拓展术语。通常决策者只采用本原术语对决策目标进行评价,而拓展术语只出现在语言变量的运算和方案排序过程中。
定义1:
其运算法则如下:
英文学习
1. Interval probability
Definition 1: n real number intervals [Li, Ui] (i=1, 2,..., n), if it is satisfied: 0≤Li≤Ui≤1 (i=1, 2,...,n), it can be used Define the corresponding probability of the basic event in Ω, call it n-dimensional interval probability, abbreviated as n-PRI, for convenience, introduce a vector: L=(L1, L2,..., Ln)T, U=(U1, U2 ,...,Un)T, then n-PRI can be recorded as n-PRI(L, U).
Definition 2:
Given a n-PRI (L, U), if there is a set of positive real numbers p1, p2,..., pn, and there are
It is said that n-PRI (L, U) is reasonable; otherwise, n-PRI (L, U) is unreasonable.
Lemma 1: An n-PRI (L, U) is reasonable if and only if
If n-PRI(L, D) is reasonable, the probability interval [Li, Ui] (i=1, 2,..., n) can be further refined, and the probability interval is:
Of course, an inaccurate n-dimensional probability interval can be refined by the above formula.
example:
A 4-dimensional probability interval {[0.1, 0.2], [0.2, 0.7], [0.3, 0.4], [0.1, 0.5]} is inaccurate, and it can be refined by the above formula as: {[0.1, 0.2 ], [0.2, 0.5], [03, 0.4], [0.1, 0.4]}.
2. Uncertain language variables
When making a qualitative evaluation of things, decision makers generally need to specify a language evaluation set in advance. Let the language evaluation set be:
S={sα|α=0,1,...,L-1}, where sα represents a language variable and L is an odd number. S has the following properties:
In practice, L generally takes 3, 5, 7, 9 and so on. It is expressed as:
In order to reduce the loss of language decision information in the process of calculation, the original discrete language evaluation set S=(s0, s1,..., sL-1) is expanded into a continuous language evaluation set S={sα|α∈[0, q]}, where q is a sufficiently large real number. If sα∈S, we call sα the original term; otherwise, we call sα an extended term. Generally, decision-makers only use the original terminology to evaluate the decision-making goals, while the extended terminology only appears in the process of linguistic variable calculation and program ranking.
Definition 1:
The algorithm is as follows:
英文翻译:谷歌翻译
参考资料:
[1]王飞,马裕超,张辉,王丹.基于区间概率优势决策模型的滑坡治理方案优选研究[J].中国安全生产科学技术,2019,15(08):88-93.
[2]刘培德,王娅姿.一种属性权重未知的区间概率风险型混合多属性决策方法[J].控制与决策,2012,27(02):276-280.
[3] Zhou H , Wang J Q , Zhang H Y . Grey stochastic multi-criteria decision-making based on regret theory and TOPSIS[J]. International Journal of Machine Learning and Cybernetics, 2015, 8(2):1-14.
本文由LearningYard学苑原创,仅代表个人观点,如有侵权请联系删除