reward_belief: Return rewards over time

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

Description

Return list of rewards over time given past actions and table of belief states

Usage

reward_belief(p0, pm, d0, dm, ds, V, Cm, Cs,state_posterior, act, disc = 0.95)

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_posterior	Table of belief states over time corresponding the actions. Can be computed using compute_belief_list
act	Vector of the past actions. Values in c('Manage', 'Survey', 'Stop')
disc	Discount factor used to compute the policy (default 0.95)

Value

Vector of rewards over time

Author(s)

Luz Pascal

Examples

## Not run: 
#values for Sumatran tiger
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

act <- c("Manage", "Survey", "Stop")
obs <- c("Seen", "Seen", "Not seen")
state_prior <- c(1,0)

state_posterior<-smsPOMDP::compute_belief_list(p0, pm, d0, dm, ds, V, Cm, Cs, state_prior, act, obs)

rew_tab<-smsPOMDP::reward_belief(p0, pm, d0, dm, ds, V, Cm, Cs,state_posterior, act)

## End(Not run)

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