/*
* static Func<StateForUtilityMaximization, alglib.ndimensional_fvec> maximize_utility_func = (state) =>
* {
* alglib.ndimensional_fvec d = (xs, fi, obj) =>
* {
* var valuation = ValuePortfolio(state, xs);
* fi[0] = -valuation.Utility;
* fi[1] = Enumerable.Sum(xs) - 1;
*
* };
* return d;
* };
*
* static Func<StateForPortfolioTargeting, alglib.ndimensional_fvec> target_portfolio_func = (state) =>
* {
* alglib.ndimensional_fvec f = (xs, fi, obj) =>
* {
* var valuation = ValuePortfolio(state, xs);
* fi[0] = valuation.Error;
* fi[1] = Enumerable.Sum(xs) - 1;
* };
* return f;
* };
*/
public static ValuationResult Optimize(OptimizationParameters state, double[] initialValues2)
{
int n = state.CurrentStocksAllocation.Length;
var weights = VectorOp.createWith(n, 1.0 / n);
double[] initialValues = VectorOp.DotDivision(VectorOp.multiplication(weights, state.StocksValue()), state.MarketData.AvgPrices);
alglib.ndimensional_fvec functionToOptimize = null;
int equalities = 0;
int inequalities = 0;
var stateType = state.GetType();
if (stateType == typeof(ReturnTargetingParameters))
{
functionToOptimize = target_return_func((ReturnTargetingParameters)state);
equalities = 5;
inequalities = initialValues.Length;
}
else if (false)//stateType == typeof(StateForUtilityMaximization))
{
//functionToOptimize = target_return_func((StateForReturnTargeting)state);
//equalities = 1;
//inequalities = 0;
}
else if (false) //stateType == typeof(StateForPortfolioTargeting))
{
//functionToOptimize = target_portfolio_func((StateForPortfolioTargeting)state);
//equalities = 1;
//inequalities = 0;
}
double[] s = new double[n];
for (int i = 0; i < n; i++)
{
s[i] = 1;// state.AvgPrices.Average() / state.AvgPrices[i];
}
double epsx = 0.00001;
double diffstep = 0.1;
double radius = 1;
double rho = 100000;
int maxits = 0;
alglib.minnsstate optimizationState;
alglib.minnsreport rep;
double[] newStocksAllocation;
alglib.minnscreatef(n, initialValues, diffstep, out optimizationState);
alglib.minnssetalgoags(optimizationState, radius, rho);
alglib.minnssetcond(optimizationState, epsx, maxits);
alglib.minnssetscale(optimizationState, s);
alglib.minnssetnlc(optimizationState, equalities, inequalities);
alglib.minnsoptimize(optimizationState, functionToOptimize, null, null);
alglib.minnsresults(optimizationState, out newStocksAllocation, out rep);
var valuation = RebalancingValuation.ValuePortfolio(state.MarketData, state.CurrentStocksAllocation, newStocksAllocation);
Console.WriteLine("{0}", alglib.ap.format(newStocksAllocation, 3));
Console.WriteLine("Expected return: {0}", valuation.ExpectedReturn);
Console.WriteLine("StdDev: {0}", valuation.StdDev);
Console.WriteLine("buysell: {0}", valuation.RebalancingCost.SharesBuySell);
//Console.WriteLine("Error: {0}", state.targ);
return(valuation);
}
