A10 <-
structure(c(0.674716681991817, -0.441716621026635, 0.622713571882636, 
2.12459319704959, -2.2200164111581, -3.61488000399049, 0.253305195789315, 
0.0766313802382238, -0.990373523500256, -0.886289788805003), .Names = c("MMM", 
"ACE", "ABT", "ANF", "ADBE", "AMD", "AES", "AET", "AFL", "A"))
A10s <-
structure(c(-0.361488000399049, -0.22200164111581, -0.875330227135, 
0.0674716681991817, 0.153262760476448, -0.0990373523500256, -0.0886289788805003, 
0.212459319704959, 0.355828337897336, 0.462271357188264), .Names = c("MMM", 
"ACE", "ABT", "ANF", "ADBE", "AMD", "AES", "AET", "AFL", "A"))
Fit10 <-
function (x, full=FALSE) 
{
	n <- length(x) / 2
	ord <- order(x[1:n])
	w <- x[1:n + n]
	w[ord <= 5] <- 0
	if(full) {
		list(weights=w/sum(w), objective=U10(w))
	} else {
		U10(w)
	}
}
Nfit10 <-
function (x, full=FALSE) 
{
	n <- length(x) / 2
	ord <- order(x[1:n])
	w <- x[1:n + n]
	w[ord <= 5] <- 0
	w[w < 0] <- 0
	if(all(w <= 0)) return( Inf )
	if(full) {
		list(weights=w/sum(w), objective=-U10(w))
	} else {
		-U10(w)
	}
}
Nfit10s <-
function (x, full=FALSE) 
{
	n <- length(x) / 2
	ord <- order(x[1:n])
	w <- x[1:n + n]
	w[ord <= 5] <- 0
	w[w < 0] <- 0
	if(all(w <= 0)) return( Inf )
	if(full) {
		list(weights=w/sum(w), objective=-U10s(w))
	} else {
		-U10s(w)
	}
}
P.lacking10a <-
function (filename = "lacking10a.png") 
{
	if(length(filename)) {
		png(file=filename, width=512)
		par(mar=c(5,4.5, 0, 2) + .1)
	}

	lack <- list(genopt=testres.genopt.10a[,2] - testres.pprobe.10a[1,2], 
		ga=testres.ga.10a[,2] - testres.pprobe.10a[1,2],
		TAopt=testres.taopt.10a[,2] - testres.pprobe.10a[1,2],
		LSopt=testres.lsopt.10a[,2] - testres.pprobe.10a[1,2],
		DEopt=testres.deopt.10a[,2] - testres.pprobe.10a[1,2],
		SANN=testres.sann.10a[,2] - testres.pprobe.10a[1,2],
		GenSA=testres.gensa.10a[,2] - testres.pprobe.10a[1,2],
		genoud=testres.genoud.10a[,2] - testres.pprobe.10a[1,2],
		soma=testres.soma.10a[,2] - testres.pprobe.10a[1,2],
		malsch=testres.malsch.10a[,2] - testres.pprobe.10a[1,2],
		DEoptim=testres.deoptim.10a[,2] - testres.pprobe.10a[1,2])
	lord <- order(sapply(lack, median))
	boxplot(lack[lord], col="gold", xlab="Deficiency", horizontal=TRUE,
		las=1)

	if(length(filename)) {
		dev.off()
	}
}
pp.denplot <-
function (..., col=c("black", "steelblue", "gold", "forestgreen", "red", "gray")) 
{
	dots <- list(...)
	dens <- lapply(dots, density)
	dx <- do.call("cbind", lapply(dens, function(z) z$x))
	dy <- do.call("cbind", lapply(dens, function(z) z$y))
	matplot(dx, dy, type="l", lty=1, col=col, lwd=2)
}
test.deopt10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			DEopt(Nfit10, algo=list(
				printDetail=FALSE, printBar=FALSE,
				min=rep(0,20), max=rep(1, 20)
			)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$OFvalue
	}
	cbind(time=time, objective=objective)
}
test.deopt10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			DEopt(Nfit10s, algo=list(
				printDetail=FALSE, printBar=FALSE,
				min=rep(0,20), max=rep(1, 20)
			)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$OFvalue
	}
	cbind(time=time, objective=objective)
}
test.deoptim10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			DEoptim(fn = Nfit10, lower = rep(0, 20), upper = rep(1, 20), 
			control=DEoptim.control(trace=FALSE, itermax=100)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$optim$bestval
	}
	cbind(time=time, objective=objective)
}
test.deoptim10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			DEoptim(fn = Nfit10s, lower = rep(0, 20), upper = rep(1, 20), 
			control=DEoptim.control(trace=FALSE, itermax=100)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$optim$bestval
	}
	cbind(time=time, objective=objective)
}
test.ga10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			ga(type = "real", fitness = Fit10, min = rep(0, 20), 
			max = rep(1, 20), monitor = NULL))
		time[i] <- sum(ttim[1:2])
		objective[i] <- tres@fitnessValue
	}
	cbind(time=time, objective=objective)
}
test.ga10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			ga(type = "real", fitness = Fit10s, min = rep(0, 20), 
			max = rep(1, 20), monitor = NULL))
		time[i] <- sum(ttim[1:2])
		objective[i] <- tres@fitnessValue
	}
	cbind(time=time, objective=objective)
}
test.genopt10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			genopt(fun = Nfit10, population = matrix(runif(20 * 20), 20), 
    			lower = 0, upper = 1, births = 2500, trace=FALSE))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$objective[1]
	}
	cbind(time=time, objective=objective)
}
test.genopt10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			genopt(fun = Nfit10s, population = matrix(runif(20 * 20), 20), 
    			lower = 0, upper = 1, births = 2500, trace=FALSE))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$objective[1]
	}
	cbind(time=time, objective=objective)
}
test.genoud10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			genoud(fn = Nfit10, nvar = 20, print.level=0, pop.size=20,
			max.generations=20,
    			Domain=cbind(rep(0,20), 1), boundary.enforce=2))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$value
	}
	cbind(time=time, objective=objective)
}
test.genoud10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			genoud(fn = Nfit10s, nvar = 20, print.level=0, pop.size=20,
			max.generations=20,
    			Domain=cbind(rep(0,20), 1), boundary.enforce=2))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$value
	}
	cbind(time=time, objective=objective)
}
test.gensa10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			GenSA(fn = Nfit10, lower=rep(0,20), upper=rep(1,20),
				control=list(maxit=300)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$value
	}
	cbind(time=time, objective=objective)
}
test.gensa10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			GenSA(fn = Nfit10s, lower=rep(0,20), upper=rep(1,20),
				control=list(maxit=300)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$value
	}
	cbind(time=time, objective=objective)
}
test.lsopt10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			LSopt(Nfit10, algo=list(
				x0=runif(20), printDetail=FALSE, printBar=FALSE,
				nS=15000,
				neighbour=function(x) x + runif(length(x))
			)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$OFvalue
	}
	cbind(time=time, objective=objective)
}
test.lsopt10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			LSopt(Nfit10s, algo=list(
				x0=runif(20), printDetail=FALSE, printBar=FALSE,
				nS=15000,
				neighbour=function(x) x + runif(length(x))
			)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$OFvalue
	}
	cbind(time=time, objective=objective)
}
test.malsch10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			malschains(fn = Nfit10,
    			lower = rep(0, 20), upper = rep(1,20), maxEvals=1.5e4, 
			trace=FALSE))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$fitness
	}
	cbind(time=time, objective=objective)
}
test.malsch10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			malschains(fn = Nfit10s,
    			lower = rep(0, 20), upper = rep(1,20), maxEvals=1.5e4, 
			trace=FALSE))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$fitness
	}
	cbind(time=time, objective=objective)
}
test.pprobe10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			trade.optimizer(prices = jjprice, variance = V10, 
			expected.return = A10, gross = 1e+06 + c(-0.5, 0.5), 
			long.only = TRUE, port.size = 5, utility = "mean-variance",
			seed=NULL))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$results["objective"]
	}
	cbind(time=time, objective=objective)
}
test.pprobe10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			trade.optimizer(prices = jjprice, variance = V10, 
			expected.return = A10s, gross = 1e+06 + c(-0.5, 0.5), 
			long.only = TRUE, port.size = 5, utility = "mean-variance",
			seed=NULL))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$results["objective"]
	}
	cbind(time=time, objective=objective)
}
test.sann10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			optim(fn = Nfit10, par = runif(20), method="SANN",
    			control=list(maxit=2e4)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$value[1]
	}
	cbind(time=time, objective=objective)
}
test.sann10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			optim(fn = Nfit10s, par = runif(20), method="SANN",
    			control=list(maxit=2e4)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$value[1]
	}
	cbind(time=time, objective=objective)
}
test.soma10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			soma(cost = Nfit10, bounds=list(min=rep(0,20), max=rep(1,20)), 
    				options=list(nMigrations=60)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$cost[tres$leader]
	}
	cbind(time=time, objective=objective)
}
test.soma10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			soma(cost = Nfit10s, bounds=list(min=rep(0,20), max=rep(1,20)), 
    				options=list(nMigrations=60)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$cost[tres$leader]
	}
	cbind(time=time, objective=objective)
}
test.taopt10a <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			TAopt(Nfit10, algo=list(
				x0=runif(20), printDetail=FALSE, printBar=FALSE,
				nS=1500,
				neighbour=function(x) x + runif(length(x))
			)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$OFvalue
	}
	cbind(time=time, objective=objective)
}
test.taopt10s <-
function (trials=100) 
{
	time <- objective <- rep(NA, trials)
	for(i in 1:trials) {
		ttim <- system.time(tres <-
			TAopt(Nfit10s, algo=list(
				x0=runif(20), printDetail=FALSE, printBar=FALSE,
				nS=1500,
				neighbour=function(x) x + runif(length(x))
			)))
		time[i] <- sum(ttim[1:2])
		objective[i] <- -tres$OFvalue
	}
	cbind(time=time, objective=objective)
}
U10 <-
function (w) 
{
	w <- w / sum(w)
	sum(w * A10) - drop(w %*% V10 %*% w)
}
U10s <-
function (w) 
{
	w <- w / sum(w)
	sum(w * A10s) - drop(w %*% V10 %*% w)
}
V10 <-
structure(c(2.10971431256697, 0.897757948573287, 0.566181049143755, 
1.50011719304549, 1.12044478464688, 1.69082908434874, 1.58090578300758, 
0.956865349017686, 1.76730730102965, 1.47174850522898, 0.897757948573287, 
1.93792106877032, 0.520542508238327, 1.4303621745831, 1.32683106568222, 
1.47938005758826, 1.40000278777414, 0.966550472224708, 1.54829237465794, 
1.26892130401874, 0.566181049143755, 0.520542508238327, 1.35629703931921, 
0.989376767113976, 0.716523791725757, 0.905374661305781, 0.949722918242164, 
0.703360008379208, 0.98741751248164, 0.852191745248497, 1.50011719304549, 
1.4303621745831, 0.989376767113976, 7.93518261961667, 2.93460873539754, 
3.27240520099923, 2.71773139243512, 1.66476650440305, 3.06729241197499, 
2.44671724214915, 1.12044478464688, 1.32683106568222, 0.716523791725757, 
2.93460873539754, 6.61357359514527, 2.21566247402706, 2.26330312256283, 
1.58144128846032, 2.36393995337297, 2.12210071189485, 1.69082908434874, 
1.47938005758826, 0.905374661305781, 3.27240520099923, 2.21566247402706, 
7.44642407634864, 3.02434819939931, 1.83447744928282, 3.23887963009818, 
3.03178327058812, 1.58090578300758, 1.40000278777414, 0.949722918242164, 
2.71773139243512, 2.26330312256283, 3.02434819939931, 4.95410681125333, 
1.70345589181136, 2.95271338981022, 2.46065981456743, 0.956865349017686, 
0.966550472224708, 0.703360008379208, 1.66476650440305, 1.58144128846032, 
1.83447744928282, 1.70345589181136, 3.0090503854516, 1.80662227565109, 
1.6560255511543, 1.76730730102965, 1.54829237465794, 0.98741751248164, 
3.06729241197499, 2.36393995337297, 3.23887963009818, 2.95271338981022, 
1.80662227565109, 4.76524713449497, 2.5398661728646, 1.47174850522898, 
1.26892130401874, 0.852191745248497, 2.44671724214915, 2.12210071189485, 
3.03178327058812, 2.46065981456743, 1.6560255511543, 2.5398661728646, 
3.90812971764787), .Dim = c(10L, 10L), .Dimnames = list(c("MMM", 
"ACE", "ABT", "ANF", "ADBE", "AMD", "AES", "AET", "AFL", "A"), 
    c("MMM", "ACE", "ABT", "ANF", "ADBE", "AMD", "AES", "AET", 
    "AFL", "A")))