Assuming that the \(z_{ijp}\) are continuous variables, you cannot get precisely what you want. What is attainable is \[ c_{ip}=\begin{cases} 0 & \textrm{if }\sum_{j=1}^{s_{max}}z_{ijp}=0\\ 1 & \textrm{if }\sum_{j=1}^{s_{max}}z_{ijp}\ge\epsilon \end{cases} \]for some small \(\epsilon \gt 0\). This is a variation on a commonly asked question; try searching the forum to find an answer. answered 25 Aug '15, 15:29 Paul Rubin ♦♦ The matrix Z is binary valued, i.e., all the elements of Z matrix are either zero or one.
(25 Aug '15, 15:37)
UbaAbd
In that case, it's fairly trivial: \(c_{ip}\ge z_{ijp}\,\forall j\) and \(c_{ip}\le \sum_j z_{ijp}\).
(25 Aug '15, 15:42)
Paul Rubin ♦♦
Thats perfect. Thanks @paul.
(26 Aug '15, 05:02)
UbaAbd

LaTeX is indeed supported: see https://www.orexchange.org/questions/9930/whatformatcouldipostequationsinorexchange.