Costs

[ArmineA]

HI,
Could you please clarify, what is the difference between Reg_ACost and Reg_wobj?

[Antti-L]

Reg_ACost reports annualized undiscounted costs, while Reg_wobj reports objective function components (which are all discounted).

You can find such information from the documentation:

Reg_ACost – Regional total annualized costs by period:
Total annualized costs in region (r) by period (t) and cost category. The cost categories are INV, INVX, FIX,
FIXX, VAR, VARX, IRE, ELS and DAM (see Table 16 below for more information).

Reg_wobj – Regional total discounted system cost by component:
Discounted objective value (EQ_OBJ) for each region (r), by cost type (uc_n) and currency ©. The cost
types are: INV, INVX, FIX, FIXX, VAR, VARX, ELS, DAM (see Table 16 below for more information).

[ArmineA]

Reg_ACost reports annualized undiscounted costs, while Reg_wobj reports objective function components (which are all discounted).

You can find such information from the documentation:

Reg_ACost – Regional total annualized costs by period:
Total annualized costs in region ® by period (t) and cost category. The cost categories are INV, INVX, FIX,
FIXX, VAR, VARX, IRE, ELS and DAM (see Table 16 below for more information).

Reg_wobj – Regional total discounted system cost by component:
Discounted objective value (EQ_OBJ) for each region ®, by cost type (uc_n) and currency ©. The cost
types are: INV, INVX, FIX, FIXX, VAR, VARX, ELS, DAM (see Table 16 below for more information).

Thanks!
How can I get Reg_wobj (Regional total discounted system cost by component) by period?

[Antti-L]

Indeed, the annual costs reported are of course in "current values" i.e. undiscounted, because that's what most users want to see.  But you can easily discount them by yourself in Excel, using the Time_NPV values reported, should you for some reason wish to do so. To get most consistent results for the period-specific discounted costs from such an exercise, you are advised to use the switches $SET ANNCOST LEV and $SET OBLONG YES.

For more info, see some earlier discussion for example here and here.

[ArmineA]

Thank you very much!
And one more question please What kind of investment include LUMPINV (Cap_New)?

[Antti-L]

> What kind of investment include LUMPINV (Cap_New)?

Investment does not include LUMPINV. LUMPINV is a qualifier related to a component of the Cap_New reporting attribute of TIMES. The purpose of that reporting component is to give "shortcut" information about the lump-sum equivalents of the investment costs, which may in literature be referred to as the full investment cost (as opposed to overnight costs, which usually excludes IDC and risk premiums). It is just complimentary reporting; users could easily calculate these lump-sum values also by themselves.

For some related discussion, you could also take a look at Cap_new discounting.

[ArmineA]

Indeed, the annual costs reported are of course in "current values" i.e. undiscounted, because that's what most users want to see.  But you can easily discount them by yourself in Excel, using the Time_NPV values reported, should you for some reason wish to do so. To get most consistent results for the period-specific discounted costs from such an exercise, you are advised to use the switches $SET ANNCOST LEV and $SET OBLONG YES.

For more info, see some earlier discussion for example here and here.

Hi Antti-L,
Could you please help?
I used switches $SET ANNCOST LEV and $SET OBLONG YES as you advised.
I used a new scenario file for switches.
But, as you can see in the attached files, the results are not changed.
Thanks a lot.

[Antti-L]

Which results are those? Can you not show the table with all dimensions expanded?
And can you show Time_NPV showing the UC_N dimension?

[Antti-L]

Ahh... but I see now that you are using RFCmd_DD, which I guess sets the switches in a DD file only! Setting GAMS control variables in a DD file are not visible elsewhere.

You should of course set the switches in the RUN file, to make them effective in the model run, and not just in the DD file. If you don't know how to set these switches in the RUN file, please ask the VEDA support.  I am myself not a VEDA expert.

[ArmineA]

Ahh... but I see now that you are using RFCmd_DD, which I guess sets the switches in a DD file only! Setting GAMS control variables in a DD file are not visible elsewhere.

You should of course set the switches in the RUN file, to make them effective in the model run, and not just in the DD file. If you don't know how to set these switches in the RUN file, please ask the VEDA support.  I am myself not a VEDA expert.

Thanks

[Antti-L]

Sorry: I checked from Amit, and he said that RFCmd_DD is written into the RUN file.  So I was wrong.

However, any switches should be written into the FLAGS section, with RFCmd_F (for example OBLONG does not work if it is in a later section).  Other than that, I have no idea why the switches do not work for you.  They all work well for me. So, very sorry again. 

[ArmineA]

Sorry: I checked from Amit, and he said that RFCmd_DD is written into the RUN file.  So I was wrong.

However, any switches should be written into the FLAGS section, with RFCmd_F (for example OBLONG does not work if it is in a later section).  Other than that, I have no idea why the switches do not work for you.  They all work well for me. So, very sorry again. 

Thanks!!

[Gary]

Hi Armine/Antti,

Let me work this out with Antti/Amit & we'll update you once sorted out.

Gary

[Antti-L]

Hi ArmineA:

I got from Gary the VD results file for the Armenia model Ref case.

> How can I get Reg_wobj (Regional total discounted system cost by component) by period?

As mentioned above, for doing that you need to use the following switches:
  ● $SET OBLONG YES
  ● $SET ANNCOST LEV

The run file I also got from Gary did indeed show that both of these were included, and so the straightforward calculation was extremely accurate. The proportional error between the Objz value and the sum of the calculated period-wise total discounted system cost by component was < 0.000000001%.

You can see the calculation in the attached Excel file, which is using the results directly from the VD results file provided by Gary.

[Gary]

Hi ArmineA:

I got from Gary the VD results file for the Armenia model Ref case.

> How can I get Reg_wobj (Regional total discounted system cost by component) by period?

As mentioned above, for doing that you need to use the following switches:
  ● $SET OBLONG YES
  ● $SET ANNCOST LEV

The run file I also got from Gary did indeed show that both of these were included, and so the straightforward calculation was extremely accurate. The proportional error between the Objz value and the sum of the calculated period-wise total discounted system cost by component was < 0.000000001%.

You can see the calculation in the attached Excel file, which is using the results directly from the VD results file provided by Gary.

Hi Antti,

Much thanks for seeing this thru to its successful resolution.