Skip to contents

Run a management strategy evaluation

Source: R/10-run_mse.R
run_mse.Rd

Runs a forward projecting MSE. Main assumptions are the projected selectivity/catchability, foraging days, and weight-at-age are the same as the terminal year of the hindcast in the operating model. Assumes survey sd is same as average across historic time series, while comp data sample size is same as last year. No implementation error and no observation error for catch!

Usage

run_mse(
  om,
  em,
  nsim = 10,
  start_sim = 1,
  assessment_period = 1,
  sampling_period = 1,
  simulate_data = TRUE,
  regenerate_past = FALSE,
  sample_rec = TRUE,
  rec_trend = 0,
  fut_sample = 1,
  cap = NULL,
  catch_mult = NULL,
  seed = 666,
  regenerate_seed = seed,
  loopnum = 1,
  file = NULL,
  dir = NULL,
  timeout = 999,
  endyr = NA,
  cores = NULL
)

Arguments

om

CEATTLE model object exported from Rceattle

em

CEATTLE model object exported from Rceattle

nsim

Number of simulations to run (default 10)

start_sim

First simulation number to start at. Useful if the code stops at specific seed/sim (default = 1).

assessment_period

Period of years that each assessment is taken

sampling_period

Period of years data sampling is conducted. Single value or vector the same length as the number of fleets.

simulate_data

Include simulated random error proportional to that estimated/provided for the data from the OM.

regenerate_past

Refits the EM to historical/conditioning data prior to the MSE, where the data are generated from the OM with simulate_data = TRUE or without simulate_data = FALSE sampling error.

sample_rec

Include resampled recruitment deviates from the"hindcast" in the projection of the OM. Resampled deviates are used rather than sampling from N(0, sigmaR) because initial deviates bias R0 low. If false, uses mean of recruitment deviates.

rec_trend

Linear increase or decrease in mean recruitment from endyr to projyr. This is the terminal multiplier mean rec * (1 + (rec_trend/projection years) * 1:projection years). Can be of length 1 or of length nspp. If length 1, all species get the same trend.

fut_sample

future sampling effort relative to last year. Log_sd * 1 / fut_sample for index and Sample_size * fut_sample for comps

cap

A cap on the catch in the projection. Can be a single number applied to all species (proportional to recommended catch) or vector of length nspp applied to each species. Default = NULL

catch_mult

A multiplier for the catch in the projection. Can be a single number or vector of length nspp. Default = NULL

seed

seed for the simulation

regenerate_seed

seed for regenerating data

loopnum

number of times to re-start optimization (where loopnum=3 sometimes achieves a lower final gradient than loopnum=1)

file

(Optional) Filename where each OM simulation with EMs will be saved. If NULL, no files are saved.

dir

(Optional) Directory where each OM simulation is saved

timeout

length of time (minutes) estimation will run before stopping a sim (default 999 minutes)

endyr

Terminal year of the MSE projection. Default = NA uses projyr from the operating model.

cores

Number of cores to use for parallel simulations. Default NULL picks parallel::detectCores() - 6, capped at 2 when running under R CMD check (which sets _R_CHECK_LIMIT_CORES_). Set to 1 to force sequential execution.

Value

A list of operating models (differ by simulated recruitment determined by nsim) and estimation models fit to each operating model (differ by terminal year).