Octave Integration

[Kenneth]

Hi Johan, 

I have a quick question regarding the integration of Yalmip in Octave. 

I've successfully installed Yalmip as per the instructions, and I am using the built in solver bmibnb to solve some global optimisation problems (see below).

When I run the problem below in Matlab, I get the expected result of -0.9999972015216918.

When I run the same problem in Yalmip using Octave, bmibnb reports: Maximum iterations or time limit exceeded (see output attached)

Is this a problem with Yalmip or simply with Octave?

Best regards,

Ken

ps: thanks for your previous help with Yalmip, which we used in our formal verification tool:

http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=7177126&tag=1

clear all

k1 = sdpvar(1,1)

k2 = sdpvar(1,1)

w1 = sdpvar(1,1)

w2 = sdpvar(1,1)

x1 = sdpvar(1,1)

x2 = sdpvar(1,1)

x3 = sdpvar(1,1)

y1 = sdpvar(1,1)

y2 = sdpvar(1,1)

y3 = sdpvar(1,1)

z1 = sdpvar(1,1)

z2 = sdpvar(1,1)

z3 = sdpvar(1,1)

F = [

0.9785660490015797<=k1<=1.0,

0.0<=k2<=0.021433950998420215,

0.9953661650749632<=w1<=1.0,

0.0<=w2<=0.0046338349250367285,

0.5382755335722992<=x1<=0.6066559085165665,

0.0<=x2<=0.0024158533507531534,

0.3930123932739281<=x3<=0.4613859614367954,

0.37756217105763773<=y1<=0.4299406061123724,

0.5700593938876276<=y2<=0.6224378289423622,

0.0<=y3<=8.833340898939976E-4,

0.20242390229414872<=z1<=0.32100205966327583,

0.0<=z2<=0.009209418740889548,

0.6776267517977816<=z3<=0.7963919294724476,

x3+x1+x2==1,

y3+y1+y2==1,

z3+z2+z1==1,

k1+k2==1,

w2+w1==1];

Rmax = -((-160*w1*y2*x1-160*w1*y2*x2+77*k1*z1*y1+160*y2*x1-77*z1*y1-77*z2*y1+160*w1*y2+160*y1)/160)

options = sdpsettings('verbose',0,'solver','bmibnb')

solvesdp(F,Rmax,options)

[Johan Löfberg]

YALMIP does not interface any nonlinear solver on octave (* Upper solver : none), hence it is not capable of generating any locally optimal solutions, unless it is lucky and a relaxations happens to be feasible.

I haven't looked at octave since I fixed the code to work there, so I don't know if there is any easily interfaced nonlinear solver available

[Johan Löfberg]

BTW, solvesdp is called optimize now