Instructions/RunInParallel.r

In vpourfaraj/lobsterCatch0.1: Models capture processes in American lobster fishery

require(bio.lobster)
require(parallel)
require(lobsterCatch)
require(devtools)
require(geosphere)
require(bio.utilities)
load_all('D:/git/lobsterCatch')

arena = matrix(0,200,200)
y=x=seq(5,195,10)
traps = expand.grid(x,y)

p = list()
p$nrowgrids = 200
p$ncolgrids = 200
p$ngrids=p$nrowgrids * p$ncolgrids
p$initlambda=.1
p$initD = 3
p$smult = 0.993
p$currentZoIInit = 1

p$trapEastStart = traps[,1]
p$trapNorthStart = traps[,2]
p$ntrapsstart = length(p$trapEastStart)

p$saturationThresholdStart = 5
p$how_closeStart = .01
p$dstepstart = 5 
p$trapSaturationStart = T

p$tSteps = 50
p$realizations = 100

smult_start = seq(.9,1,length.out=8)

plist = list()

for(i in 1:length(smult_start)){
	pp = p
	pp$smult = smult_start[[i]]
	plist[[i]]=pp
}

nCores = 8
cl <- makeCluster(nCores)
clusterEvalQ(cl,{require(lobsterCatch)
				  require(bio.lobster)
				  })
s=Sys.time()
out = parLapply(cl,plist,SimulateLobsterMovement )
Sys.time()-s

nCores = detectCores()
cl <- makeCluster(nCores)
clusterEvalQ(cl,{require(devtools)
			load_all('~/git/lobsterCatch')
			sink(paste0("~/tmp/output", Sys.getpid(), ".txt"))
				  })

out = parLapply(cl,plist, SimulateLobsterMovement)
stopCluster(cl)
saveRDS(out,'D:/Projects/LobsterCatchSimShrinkage.rds')


outM = list()
outD = list()
for(i in 1:length(smult_start)){
g = out[[i]]
g1 = lapply(g,mtr)
outM[[i]] = unlist(lapply(g1,mean))
outD[[i]] = unlist(lapply(g1,dispersion))
} 

unlist(lapply(outM,mean))
unlist(lapply(outD,mean)))
unlist(lapply(outD,mean))

#############################################################################################################
###############################################################################################################
############SPLIT VARIABLES ACROSS ITERATIONS#####################################################################
#############################################################################################################
require(bio.lobster)
require(parallel)
require(lobsterCatch)
require(devtools)
load_all('D:/git/lobsterCatch')
arena = matrix(0,200,200)
y=x=seq(5,195,10)
traps = expand.grid(x,y)

p = list()
p$nrowgrids = 200
p$ncolgrids = 200
p$ngrids=p$nrowgrids * p$ncolgrids
p$initlambda=.1
p$initD = 3
p$smult = 0.993
p$currentZoIInit = 1

p$trapEastStart = traps[,1]
p$trapNorthStart = traps[,2]
p$ntrapsstart = length(p$trapEastStart)

p$saturationThresholdStart = 5
p$how_closeStart = .01
p$dstepstart = 5 
p$trapSaturationStart = T

p$tSteps = 100
NVariation = 9
smult_start = seq(.8,1,length.out=NVariation)

nCores = detectCores()-1
plist = list()

splits = nCores/NVariation
Realizations = 21
smult_start = rep(smult_start,each=splits)
p$realizations = round(Realizations / splits)
for(i in 1:length(smult_start)){
	pp = p
	pp$smult = smult_start[[i]]
	plist[[i]]=pp
}

cl <- makeCluster(nCores)
clusterEvalQ(cl,{require(devtools)
			load_all('D:/git/lobsterCatch')
			sink(paste0("D:/tmp/output", Sys.getpid(), ".txt"))
				  })

out = parLapply(cl,plist, SimulateLobsterMovement)
stopCluster(cl)
saveRDS(out,'D:/Projects/LobsterCatchSimShrinkageFixed.rds')

mtr = function(x) apply(x$traps,2,max)

outM = list()
outD = list()

oo = list()
iu = dim_list(out)
for(i in 1:nrow(iu)){
	o = lapply(out[[i]],mtr)
	outM[[i]] = unlist(lapply(o,mean))
	outD[[i]] = unlist(lapply(o,dispersion))
	}

f = data.frame(Mean = unlist(outM),Dis = unlist(outD),Smul = rep(smult_start,each=p$realizations)