Normalization in AIMMS

34 views
Skip to first unread message

İNCİ

unread,
May 6, 2019, 11:19:01 AM5/6/19
to AIMMS - The Modeling System
Hello
i:Stands (1....1168)
j:scenario (1...50)
t:period (1...5)

I am trying to normalize the C(i,j,t). C(i,j,t) is parameter.  
I'll normalize between 0 and 1 values. İs there any way to normalize the value? 
one of the ways is to divide it into the biggest value in the series. But How we will symbolize in AIMMS?

For example;Current value

C111 =440
C211=500
C311=510
....
C116811=200

in this series. the biggest value is 510. We should normalize by dividing it by five hundred (500).
After normalization;

C111 =440/510 =0.86
C211=500/510=0.98
C311=510/510 =1
....
C116811=200/500=0.39

Best Regards

İNCİ

unread,
May 6, 2019, 11:19:01 AM5/6/19
to AIMMS - The Modeling System
Hello
i:Stands (1....1168)
j:scenario (1...50)
t:period (1...5)

I am trying to normalize the C(i,j,t). C(i,j,t) is parameter.  I'll normalize between 0 and 1 values. İs there any way to normalize the value? But How we will symbolize in AIMMS?

For example
C111 =440
C211=500
C311=510
....
C116811=200

in this series. the biggest value is 510. We should normalize by dividing it by five hundred ten (510).
After normalization
C111 =440/510 =0.86
C211=500/510=0.98
C311=510/510 =1
....
C116811=200/510=0.39

Best Regards

Mohan

unread,
May 6, 2019, 7:11:29 PM5/6/19
to AIMMS - The Modeling System
Hello,

You will need add additional indices in the sets i, j, t - say i1, j1 and t1.

Now, have another parameter C1(i, j, t) with below definition.

C(i, j, t)/$max[(i1, j1, t1), C(i, j, t)]


İNCİ ÇAĞLAYAN

unread,
May 8, 2019, 1:20:26 PM5/8/19
to ai...@googlegroups.com
Hello Mohan

Thank you so much for your reply. I added additional indices in the sets i, j, t - say i1, j1, and t1, Also parameter C1(i, j, t) with below definition. 
C(i, j, t)/$max[(i1, j1, t1), C(i1, j1, t1)]

when I write small (i=4, j=4 t=4) model, it is running. But in my big model (i=1168, j=50 t= 5), it did not run. There is no error in the model, but the model is running continuously and cannot find the result. I waited for 4 hours. Could you please say, What can be the problem? 
Best Regards

Mohan <aimms.m...@gmail.com>, 7 May 2019 Sal, 02:11 tarihinde şunu yazdı:
--
You received this message because you are subscribed to the Google Groups "AIMMS - The Modeling System" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/aimms/5539b309-bdcd-448e-9570-78e2f61f0375%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Mohan

unread,
May 8, 2019, 1:24:54 PM5/8/19
to AIMMS - The Modeling System
Hi there,

i=1168, j=50, t=5 means that you have find the max value out of 1168*50*5 = 292,000 values and as we used the max directly in the definition - this max is being computed 292,000 times. 292,000 * 292,000.

I would suggest that you create a separate parameter, say MaxValue with the below definition.

max([(i, j, t), C(i, j, t)]

Update the definition of C1(i, j, t) as below.

C(i, j, t)/$MaxValue

In this approach, you will be calculating the max value only once and storing it in the MaxValue parameter. The computation of C1(i, j, t) is a simple division.

let me know if this works or not.
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+unsubscribe@googlegroups.com.

İNCİ ÇAĞLAYAN

unread,
May 10, 2019, 1:26:25 PM5/10/19
to ai...@googlegroups.com
Hello Mohan
Thank you so much. it is working now. 
Can we find the maximum value of each vector series?
For example, İf we think that i=3, j=2 and t=2

vektor1  vektor 2 ...
C111      C121
C211      C221
C311      C321

vector 1
C111=440            
C211=500
C311=510
in this vector, the biggest value is 510. We will divide to 510 this series.

vector 2
C121=100
C221=110
C321=120
in this vector, the biggest value is 120. We will divide to 120.  İs it possible to do this?

Best Regards

Mohan <aimms.m...@gmail.com>, 8 May 2019 Çar, 20:24 tarihinde şunu yazdı:
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+un...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "AIMMS - The Modeling System" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/aimms/4aa3e0ca-868e-471a-bd24-9beb00dda81d%40googlegroups.com.

Mohan

unread,
May 10, 2019, 1:35:53 PM5/10/19
to AIMMS - The Modeling System
Hello,

Looks like you want to find the max for each vector series represented by j.

Give MaxValue an index domain j, and the definition will be the same.

MaxValue(j) with definition max[(i, k), C(i, j, k)]

This will give you MaxValue(j = 1) = 510 and MaxValue(j = 2) = 120
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+unsubscribe@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "AIMMS - The Modeling System" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+unsubscribe@googlegroups.com.

İNCİ ÇAĞLAYAN

unread,
May 12, 2019, 11:35:36 PM5/12/19
to ai...@googlegroups.com
Dear Mohan
Thank you so so much. it is running.
Best regards


Mohan <aimms.m...@gmail.com>, 10 May 2019 Cum, 20:35 tarihinde şunu yazdı:
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+un...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "AIMMS - The Modeling System" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+un...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "AIMMS - The Modeling System" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aimms+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/aimms/ce8447be-50b9-403e-99da-028e95876ac4%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages