Posts Tagged with "ISO 26262"

既に発行済みのブログであっても適宜修正・追加することがあります。
We may make changes and additions to blogs already published.
posted by sakurai on December 26, 2023 #724

台湾の読者からご質問を頂きました。原文は英語です。

論文"Generic Equations for a Probabilistic Metric for Random Hardware Failures According to ISO 26262"において、以下の2か所の式変形過程が分からないが、どうして次の式(13), (15)が導出されるのか?

図%%.1
(15) 省略、次ページで解説

当該論文の発表記事は#199です。そして式(13)の導出はブログに書かれていない部分の式変形です。そのため、今回新たにChatGPTに導出させたのでそれを示します。メールに英語で回答する都合上ChatGPTと英語で会話したので、併せて和訳を付けます。ChatGPTには元の式(13)の一部 $$ F_\text{SM}(T_\text{service})\sum_\text{i=0}^{n-1}R_\text{SM}(t-(i+1)T_\text{service}) $$ について変形させました。


ChatGPT:
The original equation is:
元の方程式は以下のとおりである。: $$ (1-R_\text{SM}(T_\text{service}))\sum_\text{i=0}^{n-1}R_\text{SM}(t-(i+1)T_\text{service}) $$ First, we apply the distributive law. This involves multiplying $1-R_\text{SM}(T_\text{service})$ with each term inside the summation. Doing so transforms the equation into:
まず、分配法則を適用する。これは$1-R_\text{SM}(T_\text{service})$を和の中の各項と乗算する。そうすることで次の式を得る。 $$ \sum_\text{i=0}^{n-1}\left[R_\text{SM}(t-(i+1)T_\text{service})-R_\text{SM}(T_\text{service}) R_\text{SM}(t-(i+1)T_\text{service})\right] $$ Next, we apply the product rule for reliability functions, which is $ R_\text{SM}(a)R_\text{SM}(b)=R_\text{SM}(a+b) $. It's important to remember that this rule applies only to terms in the form of $ R_\text{SM} $. We can apply this rule to the second term in the expanded equation:
次に、信頼性関数の積則を適用する。$R_\text{SM}(a)R_\text{SM}(b)=R_\text{SM}(a+b)$である。$R_\text{SM}$の形の項だけに適用されるので、この規則を展開された方程式の第2項に適用できる。 $$ \sum_\text{i=0}^{n-1}\left[R_\text{SM}(t-(i+1)T_\text{service})-R_\text{SM}(T_\text{service}+t- (i+1)T_\text{service})\right] $$ Upon simplifying further, the equation becomes:
さらに単純化すると方程式は次のようになる。 $$ \sum_\text{i=0}^{n-1}\left[R_\text{SM}(t-(i+1)T_\text{service})-R_\text{SM}(t-iT_\text{service})\right] $$ We can separate this into two distinct summations by splitting the terms within the summation:
我々は項を分割し2つの異なる和に分けることができる。 $$ \sum_\text{i=0}^{n-1}R_\text{SM}(t-(i+1)T_\text{service})-\sum_\text{i=0}^{n-1}R_\text{SM}(t-iT_\text{service}) $$ Now, to further transform the equation, we change the variable in the first summation from $i$ to $j=i+1$. This results in adjusting the limits of the summation as well. The transformed equation becomes:
この方程式をさらに変形するために、最初の和の変数を$i$から$j=i+1$に変更すると変換された方程式は次のようになる。 $$ \sum_\text{j=1}^{n} R_\text{SM}(t - jT_\text{service}) - \sum_\text{i=0}^{n-1} R_\text{SM}(t - iT_\text{service}) $$ We observe that for each term in the first summation, there is a corresponding term in the second summation that cancels it out, except for the terms at the boundaries. What remains are the terms for $j=n$ in the first summation and $i=0$ in the second summation. Therefore, the simplified result of the equation is:
最初の和の各項について、それを打ち消す2番目の和の対応する項があることがわかる。 残るのは、最初の和の$j=n$の項と、2番目の和の$i=0$の項である。 よって、この式を簡略化すると次のようになる。 $$ R_\text{SM}(t-nT_\text{service})-R_\text{SM}(t) $$ This represents the remaining terms after all other terms in the summations have canceled each other out.
これは、和の他の項がすべて相殺された後の残りの項を表している。


ここでこの結果を元の式(13)に代入し、$u\equiv t \bmod T_\text{service}=t-nT_\text{service}$とおけば、 $$ \begin{eqnarray} A_\text{SM}(t)&=&R_\text{SM}(t)+K_\text{SM,MPF}\left[R_\text{SM}(u)-R_\text{SM}(t)\right]\\ &=&(1-K_\text{SM,MPF})R_\text{SM}(t)+K_\text{SM,MPF}R_\text{SM}(u),\ \ \ \text{s.t. }u\equiv t \bmod T_\text{service} \end{eqnarray} $$ これにより論文(13)の最後の式が導出されました。


左矢前のブログ 次のブログ右矢


ページ: