PerformHybridOptimizationUsingSbiofitExample

Perform hybrid optimization using sbiofit
Load Data
Create a two-compartment model with infusion dose
Define the options for hybrid optimization
Fit data
Plot Results
Close Parallel Pool

Perform hybrid optimization using sbiofit

This example shows how to configure sbiofit to perform a hybrid optimization, that is, first running the optimization using the global solver particleswarm, followed by another minimization function fminunc after the global solver terminates.

Load Data

load(fullfile(matlabroot,'examples','simbio','data10_32R.mat'))
gData = groupedData(data);
gData.Properties.VariableUnits = {'','hour','milligram/liter','milligram/liter'};

Create a two-compartment model with infusion dose

pkmd                 = PKModelDesign;
pkc1                 = addCompartment(pkmd,'Central');
pkc1.DosingType      = 'Infusion';
pkc1.EliminationType = 'linear-clearance';
pkc1.HasResponseVariable = true;
pkc2                 = addCompartment(pkmd,'Peripheral');
model                = construct(pkmd);
configset            = getconfigset(model);
configset.CompileOptions.UnitConversion = true;
dose             = sbiodose('dose','TargetName','Drug_Central');
dose.StartTime   = 0;
dose.Amount      = 100;
dose.Rate        = 50;
dose.AmountUnits = 'milligram';
dose.TimeUnits   = 'hour';
dose.RateUnits   = 'milligram/hour';
responseMap = {'Drug_Central = CentralConc','Drug_Peripheral = PeripheralConc'};
paramsToEstimate   = {'log(Central)','log(Peripheral)','Q12','Cl_Central'};
estimatedParam     = estimatedInfo(paramsToEstimate,'InitialValue',[1 1 1 1]);

Define the options for hybrid optimization

rng('default');
globalMethod = 'particleswarm';
options = optimoptions(globalMethod);
hybridMethod = 'fminunc';
% Use optimset for fminsearch. Use optimoptions for fminunc and
% patternsearch.
hybridopts = optimoptions(hybridMethod);
options = optimoptions(options,'HybridFcn',{hybridMethod,hybridopts});

Fit data

unpooledFit =  sbiofit(model,gData,responseMap,estimatedParam,dose,globalMethod,...
                       options,'Pooled',false,'UseParallel',true,'ProgressPlot',true);

Plot Results

plot(unpooledFit);

Close Parallel Pool

delete(gcp('nocreate'))