RE: Output to GMPL data file
Hi Heinrich,
Thanks for your help, it works just fine.
Just had a look at the language reference too for the printf statement.
Best regards
Philippe
De : Heinrich Schuchardt
Envoyé : mardi 27 juillet 2021 20:32
À : help-glpk@gnu.org ; Philippe Jugla
Objet : Re: Output to GMPL data file
Am 27. Juli 2021 20:06:53 MESZ schrieb Philippe Jugla
:
>Hello everyone,
>
>I am looping with a .bat file to run N simulations with glpsol.
>At the end of each simulation, I would like to use the value of a variable
>which is calculated, as an input for the next simulation.
>What I have written in my .mod file doesn't work because it does not overwrite
>my storage.dat file after each simulation...
>
>in the .mod file :
>TIME = 0..T-1;
>param initial_storage;
>var storage {t in TIME};
>/* I want to output the final storage value as initial storage for next
>simulation*/
>printf ‘data; param initial_storage := ‘ & storage[T-1].val & ‘ ;’ >>
>‘storage.dat’
Use a single > here.
>> is used to append.
Have a look at print format strings too.
Best regards
Heinrich
>printf ‘end ;’ >> ‘storage.dat’
>
>the .bat file :
>for /l %%k in (1,N,1) do (
>echo solving DAY %%k...
>"%GUSEKPATH%\glpsol.exe" --cuts -m "model.mod" -d "storage.dat" -d
>"data_%%k.dat"
>)
>
>Is there a way of achieving this ?
>Thank you very much,
>
>Best regards,
>
>Philippe
>
Output to GMPL data file
Hello everyone,
I am looping with a .bat file to run N simulations with glpsol.
At the end of each simulation, I would like to use the value of a variable
which is calculated, as an input for the next simulation.
What I have written in my .mod file doesn't work because it does not overwrite
my storage.dat file after each simulation...
in the .mod file :
TIME = 0..T-1;
param initial_storage;
var storage {t in TIME};
/* I want to output the final storage value as initial storage for next
simulation*/
printf ‘data; param initial_storage := ‘ & storage[T-1].val & ‘ ;’ >>
‘storage.dat’
printf ‘end ;’ >> ‘storage.dat’
the .bat file :
for /l %%k in (1,N,1) do (
echo solving DAY %%k...
"%GUSEKPATH%\glpsol.exe" --cuts -m "model.mod" -d "storage.dat" -d
"data_%%k.dat"
)
Is there a way of achieving this ?
Thank you very much,
Best regards,
Philippe
RE: Constraint on a binary variable
Hi Heinrich That's a really elegant way to model it.. Just finished testing it, and works just fine. Thanks a lot, Best regards, Philippe De : Heinrich Schuchardt Envoyé : mercredi 30 juin 2021 17:40 À : Philippe Jugla Cc : help-glpk@gnu.org Objet : Re: Constraint on a binary variable On 6/30/21 5:24 PM, Philippe Jugla wrote: > Hi everyone, > > In my model, I have a variable "p" which has an upper bound pmax := 2.5 > and a lower bound pmin := -2.5. > I would simply like to add a binary variable "sign" which takes the values : > > 1 if variable p is positive > 0 if variable p is negative (or vice-versa) > > I thought of something like : > > s.t c1 : sign >= p[t] / pmax; > > So if p[t] is positive : sign >0 and <1, therefore sign will be forced to 1. > However, if p[t] is negative, "sign" is greater than a negative value so > the variable is no longer constrained, and can take values 0 or 1. > > How can I work things out ? > > Thanks for your help > > Best regards, > Philippe var sign, binary; var p, >= -2.5, <= 2.5; s.t. constraint: p <= 2.5 * sign; #maximize obj: p; minimize obj: p; solve; display p, sign; end; Best regards Heinrich
Constraint on a binary variable
Hi everyone, In my model, I have a variable "p" which has an upper bound pmax := 2.5 and a lower bound pmin := -2.5. I would simply like to add a binary variable "sign" which takes the values : 1 if variable p is positive 0 if variable p is negative (or vice-versa) I thought of something like : s.t c1 : sign >= p[t] / pmax; So if p[t] is positive : sign >0 and <1, therefore sign will be forced to 1. However, if p[t] is negative, "sign" is greater than a negative value so the variable is no longer constrained, and can take values 0 or 1. How can I work things out ? Thanks for your help Best regards, Philippe
Scheduling problem - cost function
Hello everyone,
I have some difficulties to add a cost parameter that depends on 2 variables to
a scheduling model in GLPK.
The problem is as follows :
If the unit has started up and been active for less than k hours than the cost
follows a certain cost function, if it has been more than k hours then the cost
follows another function.
I have tried the following example but doesn’t seem to work :
#set
set TIME:=0...T
#binary variables (already defined and working)
var start_up {TIME} binary;
var unit_on {TIME} binary;
#parameters
param T:=10;
param price {TIME};
param k:=2;
#this is what I’ve tried :
param cost {t in TIME} :=
if sum { s in TIME : s = t - k and s >= 0 } start_up[s] +
unit_on[s] < k+1 then 2.5*price[t] + 5 else 3.0*price[t] + 10;
The following error is thrown : « operand preceding < has invalid type ». I
guess it is because of the variables ? I tried to define cost as a variable but
with no success as well..
I am also trying to workaround things by defining another binary variable X
{TIME} that I would constrain, and then define the cost with something like
cost[t] = (1-X)*(cost_function_1) + X*cost_function_2 but it seems quite
complicated for this type of problem.
Does anyone know what is the issue ? How could I work things out ?
Any help or advice would be greatly appreciated !
Best regards
Philippe
RE: Constraint modelling on a subset
Hi Heinrich,
It works really well, I would have never thought of it that way.
Thanks again for your help,
Best regards
Philippe
De : Heinrich Schuchardt
Envoyé : mercredi 24 mars 2021 00:05
À : Philippe Jugla ; help-glpk@gnu.org
Objet : Re: Constraint modelling on a subset
On 23.03.21 22:28, Philippe Jugla wrote:
> Hello,
>
>
>
> I am rather new to GLPK and I am seeking help regarding the modelling of
> a constraint in a unit commitment problem.
>
> I hope someone will kindly help me on this one.
>
>
>
> I am trying to model a constraint which constrains a sum, but on a
> sequence of subsets of an initial set TIME.
>
> A bit of context with a simple example below :
>
>
>
> *#sets *
>
> set TIME := 1..T;
>
> set PLANTS :=P1, P2;
>
> * *
>
> *#parameters *
>
> param T;
>
> param max_startups_year {PLANTS};
>
> param max_startups_week {PLANTS};
>
> * *
>
> *#variable *
>
> var startup {p in PLANTS, t in TIME} binary;
>
> * *
>
> *#constraint 1 *
>
> subject to C1 {p in PLANTS}:
>
> sum {t in TIME} startup[p,t] <= max_startups_year[p];
>
> *_ _*
>
> *Now this is where I am struggling : I would like to constrain sum of
> startup[p,t] with parameter max_startups_week[p] but on subsets of the
> set TIME with step k (let’s say k=5). *
>
> * *
>
> *The following works but obviously is not flexible at all. *
>
> *It gives you the idea of what I would like to do : *
>
>
>
> sum {t in 0..5} startup[p,t] <= max_startups_week[p];
>
> sum {t in 6..10} startup[p,t] <= max_startups_week[p];
>
> …
>
> Etc…
>
> …
>
> sum {t in T-5..T} startup[p,t] <= max_startups_week[p];
>
>
>
>
> I have tried to define another set TIME_2 but it’s not satisfying as it
> is hard-coded as well…
>
>
>
> *Set TIME_2 := (0..5 union 6..10.. union /[etc]/ union T-5..T) *
>
> *subject to C2 {p in PLANTS}: *
>
> *sum {s in TIME_2} startup[p,s] <= max_startups_week[p]; *
>
>
>
> How would you work this constraint out to be robust and flexible ? At
> the end, the number of steps k should be a parameter.
>
> To simplify things, let’s say that k divides exactly set TIME.
>
>
>
> Thanks very much for your help,
>
>
>
> Philippe
>
>
All you need is an indexed constraint like:
s.t. constraint{w in WEEKS}:
max_startups_week[w] >=
sum{d in DAYS_OF_WEEK} startup[d + days_per_week * w];
Best regards
Heinrich
Constraint modelling on a subset
Hello,
I am rather new to GLPK and I am seeking help regarding the modelling of a
constraint in a unit commitment problem.
I hope someone will kindly help me on this one.
I am trying to model a constraint which constrains a sum, but on a sequence of
subsets of an initial set TIME.
A bit of context with a simple example below :
#sets
set TIME := 1..T;
set PLANTS :=P1, P2;
#parameters
param T;
param max_startups_year {PLANTS};
param max_startups_week {PLANTS};
#variable
var startup {p in PLANTS, t in TIME} binary;
#constraint 1
subject to C1 {p in PLANTS}:
sum {t in TIME} startup[p,t] <= max_startups_year[p];
Now this is where I am struggling : I would like to constrain sum of
startup[p,t] with parameter max_startups_week[p] but on subsets of the set TIME
with step k (let’s say k=5).
The following works but obviously is not flexible at all.
It gives you the idea of what I would like to do :
sum {t in 0..5} startup[p,t] <= max_startups_week[p];
sum {t in 6..10} startup[p,t] <= max_startups_week[p];
…
Etc…
…
sum {t in T-5..T} startup[p,t] <= max_startups_week[p];
I have tried to define another set TIME_2 but it’s not satisfying as it is
hard-coded as well…
Set TIME_2 := (0..5 union 6..10.. union [etc] union T-5..T)
subject to C2 {p in PLANTS}:
sum {s in TIME_2} startup[p,s] <= max_startups_week[p];
How would you work this constraint out to be robust and flexible ? At the end,
the number of steps k should be a parameter.
To simplify things, let’s say that k divides exactly set TIME.
Thanks very much for your help,
Philippe
