Genetic Algorithm stops after few iterations when constraints are added

HI

I have been working on a genetic algorithm for optimizing the calculation of power loss and volume of my converter. The algorithm works fine when no constraints are added. But once, I add the constraints the algorithm stops in a second after 3-4 iterations with the following error:

Infeasible individuals are present in the final population.

Optimization terminated: average change in the spread of Pareto solutions less than options.FunctionTolerance.

Constraints are not satisfied within constraint tolerance.

I am using GA multi-obj with mixed integers. I do understand the constraints with GA for using mixed-integers and equality constraints. But I am using inequality constraints.

dins_max = dins_min - dins_cal ;

Temp_rise_max = Temp_rise-150;

c = [dins_max,Temp_rise_max];

ceq = [];

Where I want my dins_cal >dins_min and Temp_rise<150. I have tried to workaround every method given in the GA matlab documentation. But nothing helps. and here is how i call:

fitnessFunction= @DAB_loss_volume_cal; % Function handle to the fitness function

numberOfVariables = 8; % Number of decision variables

populationSize = 100;

stallGenLimit = 100;

generations = 200;

ParetoFraction = 0.85;

lb = [10e-3, 1, 1,0.1,2e6,100e3,1,0.5e-6];

ub = [100e-3, 50, 30,0.42, 4e6,300e3,100, 5e-6];

Bound = [lb; ub]

options = gaoptimset('PopulationSize',populationSize,...

'CreationFcn', @int_pop,...

'MutationFcn', @int_mutation,...

'CrossoverFcn',@int_crossoverarithmetic,...

'StallGenLimit', stallGenLimit,...

'Generations', generations,...

'PopulationSize',populationSize,...

'ParetoFraction',ParetoFraction,...

'PopInitRange', Bound,...

'PlotFcns',@gaplotpareto);

[x, f, exitflag, output, population, score] = gamultiobj(fitnessFunction,...

numberOfVariables, [], [], [], [], lb, ub,@constraintfile, options);

variables [2,3,6,7] are integers

This is the output:

output =

struct with fields:

problemtype: 'nonlinearconstr'

rngstate: [1×1 struct]

generations: 102

funccount: 10302

message: 'Optimization terminated: average change in the spread of Pareto solutions less than options.FunctionTolerance.↵Constraints are not satisfied within constraint tolerance.'

maxconstraint: 568.3102

averagedistance: 0

spread: 0

exitflag = -2

Any help would be appreciated.

Thanks

Sneha

Answer 1

[Alan Weiss]

You cannot run gamultiobj with integer constraints. Sorry.

Alan Weiss

MATLAB mathematical toolbox documentation

[Sneha Thakur]

Thanks.