Mixed integer models for coal blending

We propose mixed integer linear programming formulation for practical blend planning at coal stockyards. The solution of this formulation is ready to be used by the stockyard operators to prepare the blend. Coal blending at stockyards are done to deliver coal of specified quality. Quality of coal is specified by 15-20 parameters like calorific value, ash content, moisture content, etc. For many of these quality parameters the resultant value after blending does not follow weighted average blend rule. Such blend rules makes linear programming inadequate for blend planning. Apart from the quality requirements, there are several limitations specific to reclaiming that need to be considered in order to practically prepare the blend. The proposed mixed integer formulation takes into account unconventional blend rules together with stockyard and reclaiming related limitations. A software is developed in which the proposed formulation is implemented. The user interfaces are demonstrated along with results.