|
|
from abc import abstractmethod |
|
|
|
|
|
class Scenario_reducer(): |
|
|
''' |
|
|
Abstract class for a scenario reducer that, |
|
|
given a fixed n, reduces an original set of scenarios |
|
|
with cardinality N to a smaller one of cardinality n. |
|
|
|
|
|
Different strategies can vary w.r.t.: |
|
|
-The metrics (Wass 1,2,inf..) |
|
|
-The selection order (Fast forward, backward, simultaneous backward) |
|
|
|
|
|
We assume a discrete uniform initial distribution on the scenarios. |
|
|
It can be generalized by appending a new value in the init of another sub-class |
|
|
|
|
|
For further details, please refer to: |
|
|
[1] Heitsch, Holger, and Werner Römisch. "Scenario reduction algorithms in stochastic programming." Computational optimization and applications 24.2-3 (2003): 187-206. |
|
|
''' |
|
|
@abstractmethod |
|
|
def __init__(self, initialSet): |
|
|
""" |
|
|
The initial set is assumed to have a dimension d x N, |
|
|
where N is the number of the initial set of scenarios |
|
|
""" |
|
|
@abstractmethod |
|
|
def reduce(self, n_scenarios: int = 1): |
|
|
""" |
|
|
reduces the initial set of scenarios |
|
|
""" |
|
|
pass |
|
|
|
