What are the optimisation techniques in MATLAB that can solve Discrete Optimisation Problems ?

Answers (1)

Mainly these two:
intlinprog - Optimization Toolbox
ga - Global Optimization Toolbox

6 Comments

Thank you John,,
Do you think both can solve a multi-objective problems?
To the extent they are applicable as an optimization. Yes. intlinprog cannot solve a nonlinear problem of course, since it is still linear programming.
As for a multi-objective problem, no tool can solve that, unless you convert the problem into a single objective. A tool like fgoalattain in the optimization toolbox is just an interface to combine multltiple objectives into one. However, it does not allow for optimization over a discrete parameter space.
So as long as you provide the scheme to reduce the multi-objective problem into a single objective one, then GA qould have no problem, as well as intlinprog IF it is a linear problem.
gamultiobj() and paretosearch() can handle multi-objective optimizations.
Thank you John & Walter...
I tried to use Paretosearch for a single objective, it is runing and gives me the result. However, the graph was empty as it shown below. Could you please help me to solve this problem, ..
This is the option I used:
options = optimoptions('paretosearch','ParetoSetChangeTolerance',1e-7,...
'PlotFcn',{'psplotbestf', 'psplotmeshsize', 'psplotbestx'},'ParetoSetSize',200);
Also I tried this:
options=optimoptions('paretosearch','ParetoSetChangeTolerance',1e-7,...
'PlotFcn',{'psplotbestf'},'ParetoSetSize',500,'Display','iter','MeshTolerance',0.01);
Yes, but gamultobj and paretosearch cannot handle discrete problems, which would seem to be of importance here.
Thanks John,
The problem is nonlinear with only one objective. I tried with intlinprog, but not working or I could not make it work. below is ga and paretosearch. ga gives me a result with graphs. However, pareto not showing the graph as explained before.
ga
options=gaoptimset('Generations',300,'CrossoverFcn',@crossovertwopoint ,'EliteCount',25,'SelectionFcn',@selectiontournament, ...
'PopulationSize',200,'FitnessScalingFcn', @fitscalingrank,'CrossoverFraction',0.5,'HybridFcn',@fmincon, ...
'PlotFcns', {@gaplotbestf, @gaplotbestindiv, @gaplotexpectation});
% [x_best,fval,exitflag,output,population,scores]=ga(@(row)massWeight(UCs,UBs,row,params),nvars,[],[],[],[],lb,ub,@(x) constraints(x,params),intcon,options);
paretosearch
options = optimoptions('paretosearch','ParetoSetChangeTolerance',1e-7,...
'PlotFcn',{'psplotbestf', 'psplotmeshsize', 'psplotbestx'},'ParetoSetSize',200);
[x_best,fval,exitflag,output,rrr]=paretosearch(@(row)massWeight(UCs,UBs,round(row),params),nvars,[],[],[],[],lb,ub,@(x) constraints(round(x),params),options);

Sign in to comment.

Categories

Products

Release

R2019b

Asked:

on 23 Jun 2020

Edited:

on 25 Jun 2020

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!