In the event of a large-scale blackout in a distribution network

a service restoration plan must be devised quickly. This paper proposes a Constraint Satisfaction Problem (CSP) based model to solve the problem. The factors considered in the service restoration plan are represented as variables

constraints and objective in terms of CSP. To speed up the online calculation

much quantitative calculation unrelated to the fault mode is performed offline. At the same time

the solution space is sorted out to improve the efficiency of problem solving so that the given objective function will be partially monotonic with respect to an individual variable. A backtracking algorithm is then employed to solve the CSP. Finally

an example result indicates the feasibility and superiority of the proposed method.