The last decade has witnessed wide-scale deployment of optical networks to support the growing data traffic. This success can be traced back to advances in optical transmission systems such as dense wavelength-division multiplexing, Raman amplification, etc., which allow a single fiber to carry several wavelengths very far, while sharing expensive equipment. However,these cutting-edge technologies require careful placement of amplifiers and other network elements to ensure error-free propagation of the signal and to minimize costs. In practice, it is common to use a set of constraints to ensure valid configurations for deployment. It is nontrivial to identify the optimal configuration under all but the simplest constraints. In this paper,we consider a set of constraints with varying flexibilities and present algorithms for efficiently computing the cost-optimal configuration under them. We also present experimental and theoretical results to evaluate the various constraints and algorithms.