Syntax error due to dimensional declaration

[CHIN JUN NAN]

Good days,

I have faced problems when I want to declare for the objective function, my R could have problems because it should be a 2-dimensional variable but I am not sure how can I declare for it in the objective function.

The ILP formulation will be:

Thanks in advance.

[AMPL Google Group]

An AMPL indexing expression must have the form "index in set-expr" where index is a name not currently defined, and set-expr is any set-valued expression. Thus, in your objective function, "P in R" is being rejected, because P is already defined (as the name of a set), and also R is not a set-valued expression (since it is the name of a parameter).

I can see that you write "P in R" because you are trying to create an AMPL expression for the following summation in your mathematical description of the objective:

But it is not standard mathematical notation to have a formula with two ϵ signs in it. Can you explain what this notation means? In particular, what values does p take when it is summed in this way?

--
Robert Fourer
am...@googlegroups.com

{#HS:1696318300-107158#}

[CHIN JUN NAN]

Dear Robert,

The notation is used to maximize the number of spatial dependencies and temporal dependencies when conducting the network monitoring. p will represent the active network flow inside a simple path; P is representing all of the possible network flows that are needed to collect all the target telemetry items requested by the application; Rsm is using to store the telemetry items that must be collected at the same node/device.

Thank you very much.

[AMPL Google Group]

You have given a general description of how p, P, and Rsm relate to your application. However, it is still not clear how the summation indexing expression -- the one with two ϵ signs -- should be interpreted mathematically.

Perhaps you can give an example. For the P and Rsm data shown in your file monitoring.dat, how many p values does

sum over, and what are those values?

--
Robert Fourer
am...@googlegroups.com

{#HS:1696318300-107158#}

On Tue, Nov 16, 2021 at 1:47 PM UTC, AMPL Modeling Language <am...@googlegroups.com> wrote:

Dear Robert,

The notation is used to maximize the number of spatial dependencies and temporal dependencies when conducting the network monitoring. p will represent the active network flow inside a simple path; P is representing all of the possible network flows that are needed to collect all the target telemetry items requested by the application; Rsm is using to store the telemetry items that must be collected at the same node/device.

Thank you very much.

On Mon, Nov 15, 2021 at 6:30 PM UTC, AMPL Google Group <am...@googlegroups.com> wrote:

An AMPL indexing expression must have the form "index in set-expr" where index is a name not currently defined, and set-expr is any set-valued expression. Thus, in your objective function, "P in R" is being rejected, because P is already defined (as the name of a set), and also R is not a set-valued expression (since it is the name of a parameter).

I can see that you write "P in R" because you are trying to create an AMPL expression for the following summation in your mathematical description of the objective:

But it is not standard mathematical notation to have a formula with two ϵ signs in it. Can you explain what this notation means? In particular, what values does p take when it is summed in this way?

--
Robert Fourer
am...@googlegroups.com

[CHIN JUN NAN]

Since I also get it from the research paper, it did not explain in detail, but the variables are given are 

, but I am not sure why suddenly become needed of p in P in Rsm...

Sorry for the confusion...

Thank you.

[AMPL Google Group]

You could consider the following AMPL statements to define the binary (zero-one) variables:

set M;
set D;
set Rs {M};
var sb {m in M, d in D, p in Rs[m]} binary;
var tb {m in M, p in Rs[m]} binary;

Then your objective function could be written like this:

maximize Objective:
   sum {m in M} sum {p in Rs[m]} sum {d in D} (sb[m,d,p] + tb[m,p]);

This still does not explain why "p in P in Rsm" appears in the paper's formulation, but maybe that was an error in printing the paper. (Unfortunately, unlike the user of a computer language such as AMPL, the author of a paper can invent any kind of strange notation or make any sort of formulation mistake without being warned by "syntax error".)

--
Robert Fourer
am...@googlegroups.com

{#HS:1696318300-107158#}

On Thu, Nov 18, 2021 at 3:04 PM UTC, AMPL Modeling Language <am...@googlegroups.com> wrote:

Since I also get it from the research paper, it did not explain in detail, but the variables are given are
, but I am not sure why suddenly become needed of p in P in Rsm...

Sorry for the confusion...
Thank you.

On Wed, Nov 17, 2021 at 6:17 PM UTC, AMPL Google Group <am...@googlegroups.com> wrote:

You have given a general description of how p, P, and Rsm relate to your application. However, it is still not clear how the summation indexing expression -- the one with two ϵ signs -- should be interpreted mathematically.

Perhaps you can give an example. For the P and Rsm data shown in your file monitoring.dat, how many p values does

sum over, and what are those values?

--
Robert Fourer
am...@googlegroups.com

[CHIN JUN NAN]

Thanks, Robert but how should I set the Rs{M} in the data file?

Thank you and Regards.

[AMPL Google Group]

It is just like for Rsm in your file monitoring.dat, except now your set is called Rs and it is indexed over M. The general idea is this:

set M := M1 M2 M3 ;
set Rs[M1] := P1 P2 P3 ;
set Rs[M2] := P2 P4 ;
set Rs[M3] := P5 P3 P2 ;

But note that this is only an example, and you will have to change it to show the correct data for your problem. (There's also an example in the subsection Indexed collections of sets in the AMPL book.)

--
Robert Fourer
am...@googlegroups.com

{#HS:1696318300-107158#}

On Fri, Nov 19, 2021 at 1:41 PM UTC, AMPL Modeling Language <am...@googlegroups.com> wrote:

Thanks, Robert but how should I set the Rs{M} in the data file?

Thank you and Regards.

On Thu, Nov 18, 2021 at 4:06 PM UTC, AMPL Google Group <am...@googlegroups.com> wrote:

You could consider the following AMPL statements to define the binary (zero-one) variables:

set M;
set D;
set Rs {M};
var sb {m in M, d in D, p in Rs[m]} binary;
var tb {m in M, p in Rs[m]} binary;

Then your objective function could be written like this:

maximize Objective:
   sum {m in M} sum {p in Rs[m]} sum {d in D} (sb[m,d,p] + tb[m,p]);

This still does not explain why "p in P in Rsm" appears in the paper's formulation, but maybe that was an error in printing the paper. (Unfortunately, unlike the user of a computer language such as AMPL, the author of a paper can invent any kind of strange notation or make any sort of formulation mistake without being warned by "syntax error".)

--
Robert Fourer
am...@googlegroups.com

[CHIN JUN NAN]

Okay, Thank you for your help.