Page 1 of 1

Random sampling problem of Stochastic Programming

Posted: Sat Apr 14, 2018 4:14 am
by Hefeng Zhai
Hello, everyone
when I do a simple SP problem, when running the problem, the result show that unknown keywords sample,can anyone help?
Best regards
The code are as fllows:

y objective
Cvmax valve maximum
Positive Variables
Cv valve;

Scalars P1 "inlet pressure of liquid flow rate" / 1000 /
pho "liquid dentisy" / 1000 /
H "pump head" / 1.3 /
err "tolerance for delivery pressure" / 20 /
D "Pipe diameter" / 0.0762 /
k "pressure drop in the pipe" /9.101e-6/
W "drived power" /31.2/
eff "pump efficiency" /0.5/
r "control valve range" /0.05/
P2 outlet pressure /625/
m flowrate /8.5/

Equations OBJ "objective to be minimized"
con1 constraint 1
con2 constraint 2
con3 constraint 3
con4 constraint 4
con5 constraint 5;

OBJ.. y =e= Cvmax;
con1.. P1*pho*sqr(Cv)+pho*H*pho*sqr(Cv)-err*pho*sqr(Cv)-sqr(m)-pho*sqr(Cv)*k*m**1.84/D**5.16-P2*pho*sqr(Cv) =l= 0;
con2.. -P1*pho*sqr(Cv)-pho*H*pho*sqr(Cv)-err*pho*sqr(Cv)+sqr(m)+pho*sqr(Cv)*k*m**1.84/D**5.16+P2*pho*sqr(Cv) =l= 0;

con3.. m*H-eff*W =l= 0;
con4.. Cv-Cvmax =l= 0;
con5.. -Cv+r*Cvmax =l= 0;

Model SP / all /;

File emp / '' /;
put emp '* problem %gams.i%'/;
randvar m Uniform 5 12
randvar P2 Uniform 250 1000
sample m P2 100 method1
putclose emp;

Set scen "scenarios" / s1*s100 /;
s_m(scen) "flowrate realization by scenario"
s_Cv(scen) "Cv by scenario"
s_P2(scen) "P2 by scenario"
s_rep(scen,*) "scenario probability" / #scen.prob 0/;

Set dict / scen .scenario.''
m .randvar .s_m
P2 .randvar .s_P2
Cv .level .s_Cv
'' .opt .s_rep /;

$onecho > lindo.opt

solve SP min y use EMP scenario dict;
Option emp=lindo;
Display s_m, s_P2,s_Cv, s_rep;

Display y.l ;

Re: Random sampling problem of Stochastic Programming

Posted: Mon Apr 16, 2018 10:26 am
by Lutz

There are some smaller issues in your model (e.g. "Option emp=lindo;" should be set before your solve statement and not afterwards), but I could not reproduce that "sample" is an unknown keyword. Maybe the GAMS version you are using is too old? "Sample" as keyword was introduced with GAMS 24.1.1.

Best regards,

Re: Random sampling problem of Stochastic Programming

Posted: Sat Apr 21, 2018 10:28 am
by Hefeng Zhai
Thanks a lot, I will try it.
Best Regards~~