Theorem 9.1
If \(O\) is a bounded open set and \(p\) is a point of \(O\) then there is a unique open interval containing \(p\) which is a subset of \(O\) whose endpoints do not lie in \(O\text{.}\)
If \(O\) is a bounded open set and \(p\) is a point of \(O\) then there is a unique open interval containing \(p\) which is a subset of \(O\) whose endpoints do not lie in \(O\text{.}\)
Now that we know that the intervals described in Theorem 9.1 exist, we can formally define them and give them a name.
If \(O\) is a bounded open set and \(p \in O\) then the open interval containing \(p\) which is a subset of \(O\) and whose endpoints are not in \(O\) is called the component of \(O\) containing \(p\).
If \(O\) is a bounded open set, then the set of all components of \(O\) is a countable collection of mutually disjoint open intervals whose union is \(O\text{.}\)
If \(S\) is any interval (open, closed, or half-open), then we define \(L(S)\) to be the length of \(S\). For example, if \(S = [a,b],\) then \(L(S)= b-a\text{.}\) If \(G\) is a finite collection of mutually disjoint open intervals then \(L(G)\) denotes the sum of the lengths of the elements of \(G\text{.}\) If \(G = \{g_1, g_2, g_3, \dots \}\) is a countable collection of mutually disjoint open intervals lying in an open interval, then \(L(G) = \sum_{i=1}^\infty L(g_i)\text{.}\)
If \(G\) is a finite collection of mutually disjoint open intervals lying in the open interval \((a,b)\text{,}\) then \(L(G) \leq b-a\text{.}\)
If \(G\) is a countable collection of mutually disjoint open intervals lying in the open interval \((a,b)\) then \(L(G) \leq b-a\text{.}\)
If \(G\) is a collection of point sets then \(G^*\) denotes the set which is the union of the members of \(G\text{;}\) that is, \(G^* = \bigcup_{g \in G} g\text{.}\)
If \(G\) and \(H\) are countable collections of mutually disjoint open intervals and \(G^* \subseteq H^*\text{,}\) then \(L(G) \leq L(H)\text{.}\)
A point set \(M\) is said to be closed if and only if no point not in \(M\) is a limit point of \(M\text{.}\)
If \(O\) is an open set which is a subset of the closed interval \([a,b]\text{,}\) then \([a,b]-O\) is a closed point set, and if \(M\) is a closed point set which lies in an open interval \((a,b)\text{,}\) then \((a,b)-M\) is an open set.
The statement that the set \(G\) of open intervals properly covers the set \(M\) means that every point of \(M\) lies in a member of \(G\) and every member of \(G\) contains a point of \(M\text{.}\)
If \(M\) is a bounded point set then by the outer measure of \(M\), denoted \(m_o(M),\) is meant the greatest lower bound of the set of all \(L(G)\) where \(G\) is any collection of mutually disjoint open intervals which properly cover \(M\text{.}\)
Show that if \(M\) is a countable point set then the outer measure of \(M\) is zero.
If \(O\) is a bounded open set and \(G\) is the set of all components of \(O\text{,}\) then \(m_o(O)=L(G)\text{.}\)
If \([a,b]\) is a closed interval, then \(m_o([a,b])=b-a\text{.}\)
If \(M\) is a bounded point set and \(I\) and \(J\) are two closed intervals containing \(M\text{,}\) and \(I \subset J\text{,}\) then \(m_o(I)-m_o(I-M) = m_o(J)-m_o(J-M)\text{.}\)
If \(M\) is a bounded point set and \(I\) and \(J\) are two closed intervals containing \(M\text{,}\) then \(m_o(I)-m_o(I-M) = m_o(J)-m_o(J-M)\text{.}\)
If \(M\) is a bounded point set, then the inner measure of \(M\), denoted \(m_i(M),\) means \(m_o(I)-m_o(I-M)\) for some closed interval, \(I\text{,}\) containing \(M\text{.}\)
The statement that the point set \(M\) is measurable means that \(m_o(M) = m_i(M)\text{.}\) When \(M\) is measurable, \(m_o(M)\) is called the measure of \(M\) and denoted by \(m(M)\text{.}\)
If \(M\) is a bounded, measurable set, and \(I\) is a closed interval containing \(M\text{,}\) then \(I-M\) is measurable.
If \(H\) and \(K\) are disjoint, bounded sets, then \(m_o(H \cup K) \leq m_o(H) + m_o(K)\text{.}\)
If \(H\) and \(K\) are disjoint, bounded measurable sets then \(H \cup K\) is measurable.
If \(H\) and \(K\) are disjoint, bounded measurable sets then \(m(H \cup K) = m(H) + m(K)\text{.}\)
If \(M\) is a closed interval or an open interval, then \(M\) is measurable and \(L(M)~=~m(M)\text{.}\)
If \(G\) is a collection of open intervals covering the closed interval \([a,b]\) then there is a finite subcollection of \(G\) which also covers \([a,b]\text{.}\)