simulations_tab: Tabulate simulations

In conservation-decisions/smsPOMDP: POMDP problem solver: When to stop managing or surveying cryptic threatened species ?

Description

Function for tabulating the means and standard deviation of rewards and belief state if extant for nbSimul simulations, following the optimal policy.

Usage

simulations_tab(p0, pm, d0, dm, ds, V, Cm, Cs, state_prior, Tmax, 
                          disc = 0.95, nbSimul = 100)

Arguments

p0	Local probability of persitance : P(extant/extant, survey or stop).
pm	Local probability of persitance if manage : P(extant/extant, manage).
d0	Local probability of detection : P(present/extant, stop).
dm	Local probability of detection : P(present/extant, manage).
ds	Local probability of detection if survey : P(present/extant, survey).
V	Estimated economic value of the species ($/yr).
Cm	Estimated cost of managing ($/yr).
Cs	Estimated cost of survey ($/yr).
state_prior	Initial belief state, vector of 2 values (belief state extant and extinct), between 0 and 1.
Tmax	Horizon of the simulation. Integer
disc	Discount factor used to compute the policy (default 0.95)
nbSimul	Number of simulations to run (default 100). Integer

Value

A data.frame. Each row corresponds to a time step. The columns are:

mean_belief	Mean of the belief state extant
up_belief	Upper bound (95%) of the belief state extant
low_belief	Lower bound (95%) of the belief state extant
mean_reward	Mean of the discounted intant reward
up_reward	Upper bound (95%) of the discounted intant reward
low_reward	Lower bound (95%) of the discounted intant reward

Author(s)

Luz Pascal

Examples

 ## Not run: 
#values for Sumatran tigers
pen <- 0.1
p0 <- 1-pen
pem <- 0.05816
pm <- 1 - pem
V <- 175.133
Cm <- 18.784
Cs <- 10.840
d0 <- 0.01
dm <- 0.01
ds <- 0.78193

#Initial belief state
state_prior <- c(0.9,0.1) #extant : 0.9, extinct : 0.1

#Horizon of the simulation
Tmax <- 20

simulations_tab(p0, pm, d0, dm, ds, V, Cm, Cs, state_prior, Tmax)
 
## End(Not run)

conservation-decisions/smsPOMDP documentation built on Oct. 27, 2020, 10:44 p.m.