Natural disasters such as storms, floods, earthquakes cause a great deal of damage to human lives and properties, and this damage is aggravated by disruptions to some critical infrastructures, such as transport system, electricity and water supplies. The primary impact of natural disasters on power systems would be severe power outages. When the period of power outage is prolonged the suffering of the customers becomes far more worsened. Consequently, it is desired to restore power distribution systems promptly. In this thesis, we are concerned with a so-called repair crew scheduling problem that minimizes the total restoration time by adequately scheduling repair personnel.Given information about fault locations and estimated repair times along with travel time data between locations, we investigate mathematical optimization models, specifically two mixed-integer programming (MIP) models. The first MIP model is formulated in analogy to the identical multiple machine scheduling problem, while the second model is proposed in an effort to reduce solution times. Noting that the repair crew scheduling problem is NP-hard, a heuristic method is also proposed. A numerical study was conducted to compare computational performances of the aforementioned exact models and the heuristic method. The computational results revealed that the second MIP model is promising when time allows. It is also observed that the heuristic method can be practically implemented under time pressure. Based on this observation, we propose an implementation plan that combines the second MIP and the heuristic method in a dynamic circumstance.