Decision Variable Class
See also Decision Variable Property Reference for a detailed list of properties for this class of object.
Contents
- Decision Variables
- Key Property
- Type
- Injecting Decision Variable Terms into the Formulation
- Definition Constraint and Time Lag
- Switching in Simulation Phases
- Solution
1. Decision Variables
Decision Variable objects represent generic decision variables that are defined in the mathematical programming problem associated with the simulation. When you create a Decision Variable object a decision variable (matrix column) is allocated to each period of the simulation. You can then use these variables in Constraint equations.
Decision Variable objects are useful when you need to define a Constraint on an aspect of the simulation that is not definable with the default Constraint coefficients, or to associate a cost with a complex set of conditions. Generic Decision Variable objects may also be used to represent commodities that are not directly modelled by the default set of objects in the simulator e.g. an industrial by-product that is a function of other variables that exist in the simulation.
Other potential uses of Decision Variable are:
- Modelling commodities with a demand such as desalination as a product of electric power and heat.
- Modelling the storage electricity or any other commodity.
- Calculating values that are functions of intrinsic decision variables and passing these to the solution.
- Cancelling out default objective function coefficients by defining a decision variable to be equal to an intrinsic variable with the opposite objective function term.
- Enforcing integer restrictions on decision variables that are intrinsically linear.
- Modelling piecewise or general non-convex or discontinuous functions.
2. Key Property
The key property for Decision Variable is Objective Function Coefficient and its period type variants for hour, day, week, month and year types. The period type of this property sets the period type of the Decision Variable. One of these properties must be defined for the Decision Variable to be included in the Model and hence you can use Scenario on this property to switch a Decision Variable in/out as required.
If you define Objective Function Coefficient there will be one decision variable in the mathematical programming problem for each interval of the horizon. You can change the optimization period by instead defining one of:
- Objective Function Coefficient Hour
- Objective Function Coefficient Day
- Objective Function Coefficient Week
- Objective Function Coefficient Month
- Objective Function Coefficient Year
It is important to note that the period type of a decision variable must not be greater than that of the given simulations phase's step size.
3. Type
Decision Variables can be continuous, integer, binary, semi-continuous or semi-integer according to the Type attribute. By default the variables are non-negative with no upper bound. This can be changed using the Lower Bound and Upper Bound properties.
4. Injecting Decision Variable Terms into the Formulation
Coefficients on a Decision Variable can be included into the 'core' formulation via properties on the memberships, for example by defining Decision Variable Nodes Net Injection Definition Coefficient.
5. Definition Constraint and Time Lag
Decision Variable includes the Definition collection which is of Constraint type. The Constraint placed in this collection automatically gets a coefficient on the Decision Variable equal to the Value Coefficient (default of 1).
It is important to set both the definition constraint and the corresponding Decision Variable to be of the same period type - see Objective Function Coefficient and Constraint Period Type.
This definition constraint is not required but is important when you need to define a Decision Variable with a Time Lag - see that topic for details.
6. Switching in Simulation Phases
Decision Variable objects formulation can be controlled by simulation phase with the attributes: Include in LT Plan, Include in PASA,Include in MT Schedule and Include in ST Schedule.
If a Decision Variable is excluded from the current simulation phase, any Constraint that involves this Decision Variable is also excluded from the simulation phase.
7. Solution
Decision Variable reports the Value at the optimal solution along with the Reduced Cost.