A sampling approach to real-time performance monitoring of digital transmission systems

The author proposes a sampling-resource-allocation problem for determining the optimal sample size for each of multiple sampling activities competing for a single sampling resource. The problem is formulated as a nonlinear knapsack problem with lower and upper binding restrictions and an additive objective function. The squared coefficient of variation is proposed for the objective function. The objective function seems to be suitable for the real-time performance monitoring of an extension of the classical method of determining the minimum sample size for a single sampling activity. The sampling resource allocation problem arises from many practical application areas such as real-time performance monitoring of integrated digital networks carrying live traffic. A solution procedure is also provided.<>