The paper proposes a new approach to the mean-variance-hedging problem under transaction costs. This approach is based on the idea of dividing the gain functional into two parts. One part representing the gains resulting from a pure buying strategy, and the other part representing the gains resulting from a pure selling strategy. The problem will be studied in a general incomplete market in discrete time. Some technical assumptions such as the RAS condition are excluded.
- Sum of closed convex cones in L
- Transaction costs