Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

The Decision Module in Delft-FEWS is used to implement decision logic and evaluation for barriers. With this module we can iteratively evaluate configurated decision rules. The configuration file of the Decision Module contains the definition of one or more Decision Trees. These Decision Trees defined in the Decision Module are associated with a Barrier definition which are defined by the Barriers configuration file.

Some important prerequisites

...

Barrier states should be defined

...

When evaluating the decision logic, it is relevant to take into account the fact that the model state will change after the barrier state has been updated (changed). To evaluate the decision logic for subsequent steps in the process, this implies that we will need to re-run the model to take this state change into account. Also, if there are multiple barriers in our area of interest, this implies that if the state of one of these barriers changes, we need to update our model simulation before we can assess the decision logic for the other barrier(s).

The decision evaluation process

...

The entire process can be summarized as follows:

  1. Run a baseline simulation (forecast) taking the actual state of the barriers as a starting point.
  2. Evaluate the decision logic based on the previous baseline simulation.
  3. If relevant criteria are met, the barrier state changes. Here, we distinguish between criteria which indicate if a state change is required. And criteria which indicate when a state change is required.
  4. Run a new simulation taking the barrier state change into account.
  5. Evaluate the decision logic of this simulation. (Note that the decision logic will only be evaluated over the period following the latest barrier state change.)
  6. Loop this process starting from step 2 3 until no further state changes are relevant criteria are met and therefore no state change is required.

While the decision logic is model independent, the model should be fed with the appropriate timeseries representing the appropriate barrier states (for example timeseries of crest height, crest level and gate height for various barrier elements).

...