Testing effort control problem for an S-shaped software reliability growth model

The software testing phase generally consists of failure observation and fault removal processes. Each of these processes consumes different type of testing effort (CPU time, Manpower). The time dependent behaviour of the testing effort and its relation to Software Reliabilty Growth has been studied earlier. It was assumed that the time between the failure observation and its subsequent removal is negligible. In other words, the testing phase consisted of only one process (failure observation process). The testing effort was assumed to follow either an exponential or a Rayleigh type distribution.