Further results on optimal $ (n, \{3, 4, 5\}, \Lambda_a, 1, Q) $-OOCs

Let \begin{document}$ W = \{w_1, w_2, \cdots, w_r\} $\end{document} be a set of \begin{document}$ r $\end{document} integers greater than 1, \begin{document}$ \Lambda_a = (\lambda_a^{(1)}, \lambda_a^{(2)}, \cdots, \lambda_a^{(r)}) $\end{document} be an \begin{document}$ r $\end{document} -tuple of positive integers, \begin{document}$ \lambda_c $\end{document} be a positive integer, and \begin{document}$ Q = (q_1, q_2, \cdots, q_r) $\end{document} be an \begin{document}$ r $\end{document} -tuple of positive rational numbers whose sum is 1. Variable-weight optical orthogonal code ( \begin{document}$ (n, W, \Lambda_a, \lambda_c, Q) $\end{document} -OOC) was introduced by Yang for multimedia optical CDMA systems with multiple quality of service requirements. In this paper, tight upper bounds on the maximum code size of \begin{document}$ (n, \{3, 4, 5\}, \Lambda_a, 1, Q) $\end{document} -OOCs are obtained, and infinite classes of optimal \begin{document}$ (n, \{3, 4, 5\}, \Lambda_a, 1, Q) $\end{document} -OOCs are constructed.