Adjust upper and lower bounds for market constraint

[FelixStege]

No description provided.

[FelixStege]

The market constraints can lead to infeasible optimization problems in a real-time environment. With real-time, the full control sent by the Baseline MPC is interpolated onto the prediction horizon of the subsequently (and therefore slighly later) executed shadow mpcs, which leads to a shift of the control signal by the computation time of the Baseline. This causes linear interpolation over the collocation points for the constraints on the control variable of the shadow mpcs resulting in an infeasible problem sicne the control variable needs to be constant over the collocation interval.

[FelixStege]

The current solution fails in the case of negative values for the control variable. The parsing should take the absolute values to avoid problems