In recent years, traffic volume is increasing tremendously. In line with this increment,rntelecom operators expand their network infrastructure. This increases power consumption ofrnthe network. Especially in a backbone network, where power consumption is dependent onrntraffic volume, the increase in power consumption is becoming critical. rnResearchers are made efforts dedicated to reduce unnecessary energy waste in backbonernnetworks using physical topology optimization for highly connected networks, but thisrnoptimization is not feasible for low connected networks like ethio telecom. In addition, linernamplifiers are placed at regular interval of 80km without considering physical and economicrnrestrictions. In practice, amplifier placement is commonly implemented by operators tornguarantee a signal quality considering placement location with availability supply power,rnshelter, and physical security with the expense of high power consumption. Current ethiorntelecom line amplifiers are placed at span length ranging from 25 to 120km considering thesernphysical and economic restrictions; due to this, ethio telecom is subjected to high powerrnconsumption in line amplifiers. rnIn this thesis, optimized line amplifier placement, which takes power consumption andrnphysical and economic restrictions in to account is investigated. Mixed Integer LinearrnProgramming (MILP) formulation which takes span length and input power level constraintrnis proposed. The proposed power-saving approach is evaluated considering ethio-telecomrnNorth circle optical backbone network topology. MATLAB is used to assign optimumrnplacement location. The result is compared with theoretical standard (80km spacing) and thernpractical (ethio telecom current deployment scenario) using number of amplification sites andrnpower saving performance. The comparison results show that amplifier site placement usingrnthe new approach can minimize the number of amplification sites by 9 from the theoreticalrnapproach and by 5 from the existing configuration and a 2% power saving can be achieved inrnthe case study network portion.