Chapter 23 Averaging MPs

In some cases users may wish to provide management advice by averaging the recommendations from several different well-performing management procedures. This of course is a new management procedure in itself, and should be tested in MSE before being adopted for use in a fishery.

The makeMeanMP function can be used to create a new MP that averages the results of several MPs.

For example, suppose we wished to develop an MP that averages the results of 4 output control MPs: BK, DBSRA, Fadapt and Rcontrol.

This can be achieved by the following:

avgMP <- makeMeanMP(c("BK", "DBSRA", "Fadapt", "Rcontrol"))
class(avgMP)

## [1] "MP"

And now we can test our new MP in MSE. We will run a decent run of simulations so the results are stable:

OM <- DLMtool::testOM
OM@nsim <- 200
MSE <- runMSE(OM, MPs=c("avgMP", "BK", "DBSRA", "Fadapt", "Rcontrol"), parallel = TRUE)

## Exporting custom MPs in global environment

## Running MSE in parallel on 10 processors

## MSE completed

How did our newly created averaging MP perform?

Tplot(MSE)

##         MP PNOF  LTY P100  P50  P10 Satisificed
## 1    avgMP 0.50 0.85 0.43 0.69 0.93       FALSE
## 2       BK 0.42 0.56 0.34 0.52 0.81       FALSE
## 3    DBSRA 0.66 0.91 0.58 0.85 0.99        TRUE
## 4   Fadapt 0.50 0.67 0.46 0.64 0.83       FALSE
## 5 Rcontrol 0.91 0.37 0.88 0.95 0.99       FALSE

More information on creating your own MPs can be found in the Developing Custom Management Procedures chapter.