Entity Class
| Description: | Ownership and/or strategic entity |
| Detail: |
See also Entity Property Reference for a detailed list of properties for this class of object.
Overview
The Entity class represents a strategic interest in one or more Markets where a given Commodity is sold, and that is produced by one or more Facilities. An Entity can 'own' those facilities outright or Share them with other entities. It reports the Revenue accrued from Sales of the Commodity in those Market(s) and the Total Production Cost from Production of those facilities and calculates a Net Profit. In stochastic optimization, Entity can place constraints on the variance of that Net Profit i.e. perform 'risk-constrained' optimization.
How to Use Entity
Entity becomes an active part of the optimization in stochastic optimization i.e. when:
- The simulation is multi-sample i.e. Stochastic Risk Sample Count > 1; and
- The active simulation phase is running in stochastic mode e.g. ST Schedule Stochastic Method = "Stochastic" (note the other simulation phases have similar settings).
Stochastic mode differs from 'regular' Monte Carlo simulation in that the samples are solved simultaneously in a single optimization problem rather than in separate/independent problems, and therefore it is possible to place constraints (and objectives) that link the decisions and outcomes across the full sample set. Stochastic optimization is solved with the "scenario-wise decomposition" technique and allows the definition 'first-stage' and 'recourse' decisions with inputs such as Facility Production Non-anticipativity. The outcome of this optimization, although more robust than Monte Carlo simulation, might still show scenarios where Net Profit is unacceptable, even if the probability is low. To control the Conditional Value at Risk (CVaR) of Net Profit you can enable Entity Formulate Risk and define Entity Acceptable Risk and Risk Level as outlined in the example below.
Refer here for an explanation of VaR and CVaR.
Example
Figure 1 shows a histogram of the total Net
Profit (values are in 1000s and for a one week window) obtained
across 100 samples of a portfolio optimization. This example was
created with an Entity that owns a number of facilities and the trades
in a market for the commodity produced by those facilities. The
simulation setup has Stochastic Risk Sample Count = 100
and the Market Price
is defined by a Variable object.
Additionally, Facility Production
Non-anticipativity is set to $10,000 which prevents these
resources from having perfect foresight of the market prices.
Figure 1: Histogram of Net Profit
Querying Net Profit in the GUI Solution Viewer automatically
calculates and reports the CVaR at 1%,5% and 10%. In this case the
CVaR at 5% is $217,459. We can now take value as the Acceptable
Risk (217,459) input along with the 5%
Risk Level (0.05). If we were to rerun the simulation with
inputs we would obtain exactly the same result. The constrain the CVaR
i.e. find a more risk averse solution, we increase the Acceptable
Risk or decrease the Risk Level.
In this case we have rerun the simulation with the input shown in
Table 1.
| Property | Value | Data File | Units |
|---|---|---|---|
| Formulate Risk | Yes | Yes/No | |
| Risk Level | 0.05 | - | |
| Acceptable Risk | 250,000 | $ |
This yields the histogram of Net Profit shown in Figure 2 (the orange
bars compared to the original blue). The CVaR constrained outcome has
altered production at the facilities to reduce the incidence of low
profit outcomes, as desired and the new CVaR reported in the solution
viewer query of Net Profit is $250,000.
Figure 2: Histogram of Net Profit with CVaR Constraints