This thesis deals with optimization of flows in a network with focus on solutions based on linear programming. The theoretical part is divided into three main sections: maximal flow, minimal flow and maximal flow with minimal costs. The first part focuses on theoretical basics of maximal flow problem and mainly on these algorithms: Ford-Fulkerson, Dinic/Edmonds-Karp, Three Indians algorithm and Goldberg push-relabel algorithm. These are explained on examples. The other chapters explain the minimal flow and the maximal flow with minimal costs problems with simple algorithms for their solutions, again explained on some examples. The practical part includes software solution of maximal flow problem in application MaxTok, which can be found on ...