Creating Flight Plans from Timetable Data Testing

[tgibson]

Hi,

I now get 15 flight plans using the data above whether I click the Ensure Closure box or not; it used to create 19 plans in that case. A nice improvement.

That said, it did generate the same number of non-circular flights as did 3.2.03 - flight plans number 2,3,6,7,8,9,11. The MRAI Compiler does not generate any using the same data.

 

[gadgets]

Tom, at no point have I suggested the flight plans would be identical - only that the number of flight plans generated would be equivalent.

The content of an individual flight plans is entirely dependent on the starting point of the sequence. Change it by one minute and, if that shifts a single leg to the other side of the start point and you'll get different results. AIFp finds and uses the longest gap in the data as the starting point. MRAI always uses midnight Sunday/Monday.

I've done all I can reasonable do with AIFP. Before making any more conclusions, why don't you compare the two using data where you believe MRAI is not giving satisfactory performance - for example, for a cargo airline where the longest gap is likely to occur mid-day and not over midnight.

Don

 

[tgibson]

Thanks, I do appreciate your efforts. I will indeed compare both compilers from now on, with any plans that cause problems.

 

[gadgets]

One last try. Development release 3.2.03(d) applies "brute force" to the algorithm. As I explained yesterday, the sequence of legs is entirely dependent upon the starting point of their selection. So, now, if a flight plan doesn't close, AIFP will try every other possible starting point until it either does close or exhausts. In the latter case, AIFP "remembers" the (failed) flight plan having the longest number of legs and uses it.

On the downside, there is no guarantee that this brute force method with generate the smallest number of FPs, but they all will close - assuming the data supports full closure. So, it seems like a good idea to allow the user to select single pass (which, because of the recently introduced "lull-detection", is likely to generate the fewest FPs) or multi-pass which should guarantee full closure - again, assuming the data supports it.)

3.2.03(d) does not include such a switch, but the next release will.

Don

 

[tgibson]

Hi,

Now we're talking! This version creates many more plans with completely circular routes - great job!! When AIFP compiles the AMX plans attached above, it creates 15 flight plans, all of them circular, just like the MRAI Compiler. But it does a more efficient job of assigning plans to the first 4 (of 6) DC-3s which allows me to manually delete the two planes that have very few flight legs (12 and 6 legs/week), if desired. This can reduce the number of planes parking for long times at large airports.

Also, compiling local service airline plans (that make many multi-stop flights every day) limit the number of flight plan legs to 98 (unlike the MRAI Compiler) which allows the AI plane to appear in FS (flights of over 100 legs will cause the plane to not show up) - nice!

I have only one further request - since I have dozens of old plans that I would like to recompile, they all use the MRAI Compiler style of aircraft types (it's the 3 letter ICAO for the plane). When I load those plans, AIFP requires me to specify the aircraft used for each type of aircraft listed in the file. This is fine, but then the aircraft used are not transferred into the proper AC# slots - all plans are still using AC#0. Here is a picture of that result (using an earlier version but the result is the same with the latest):

You can replicate this result by loading the plans_amx62.txt file attached above into AIFP, and then assigning some random aircraft in your FS installation to each aircraft type. All end up as AC#0.

Since the user has already specified which aircraft should be used for each aircraft type listed in the file, I would hope that AIFP could assign those AC# and reg no. values using the same algorithms that it uses when it *does* find the aircraft in my Aircraft List? That would be so helpful, thanks.

Once again, great job.

 

[gadgets]

I have only one further request - since I have dozens of old plans that I would like to recompile, they all use the MRAI Compiler style of aircraft types (it's the 3 letter ICAO for the plane). When I load those plans, AIFP requires me to specify the aircraft used for each type of aircraft listed in the file. This is fine, but then the aircraft used are not transferred into the proper AC# slots - all plans are still using AC#0.

Currently, AIFP expects the aircraft to be assigned in the flightplan file and recoded in the AC# column. But, in the files you are loading, the aircraft number is blank and the aircraft type is in the registration field - not an intuitive arrangement.

What you are asking for MAY possible, but I'll need to investigate very carefully and get back to you.

Now we're talking! This version creates many more plans with completely circular routes - great job!! When AIFP compiles the AMX plans attached above, it creates 15 flight plans, all of them circular, just like the MRAI Compiler.

Yes, as we've already discussed, the key is the starting point. The version you have every looks for the longest lull for every aircraft type and uses the end of that lull as the starting point - maximizing the chances for circularity. MRAI, on the other hand, always starts at night Sunday/Monday. So, it is almost certain AIFP will find the same starting point (or perhaps a better one.) and will almost certainly out-perform MRAI if the schedule doesn't have a lull over midnight Sunday/Monday.

Due to some work I did yesterday, that level of uncertainty is now gone. AIFP WILL find a starting point that equals or betters MRAI. And, I believe I can even improve on that and generate the smallest possible number of FPs. What I don't know at the moments is if the time required to do so is reasonable.

Stay tuned.
Don

 

[tgibson]

Hi Don,

Great news, looking forward to any further improvements.

Thanks,

 

[gadgets]

I have only one further request - since I have dozens of old plans that I would like to recompile, they all use the MRAI Compiler style of aircraft types (it's the 3 letter ICAO for the plane). When I load those plans, AIFP requires me to specify the aircraft used for each type of aircraft listed in the file. This is fine, but then the aircraft used are not transferred into the proper AC# slots - all plans are still using AC#0.

Tom, just to be sure I understand:

• Are you planning simply to recompile from timetable data? If so, AIFP should already find all the available required aircraft in your "stable" (based on matching ATC_Types and parking codes), place them in the Aircraft List and assign them to flightplans. Or
• Do you have some flight plan text files in MRAI-compiled format (all AC#=0 with a/c type in the registration field) which you want to load into the Flight Plans List and have aircraft be assigned. If this is the case, do you:
  • have matching aircraft text files (load with Aircraft/Load New aircraft List) and you want those aircraft assigned to the FPs, or
  • want to simply load the MRAI-style FPs and have AIFP find the available aircraft (again based on matching ATC_Types and parking codes), place them into the Aircraft List and assign them to FPs.

Either can easily be done, but I need to know which you want. However, in the case of the latter, you would also have to specify parking code(s). Please respond ASAP

Don

 

[tgibson]

Here is the sequence of what happens now, and what I would like to happen:

If you look at the AMX flight plan file I posted above as text in the message, you can see that the aircraft types are listed as the ICAO aircraft types (as required by the MRAI Compiler. I don't want to have to change all the dozens of flight plans I have already in this format. We'll use this file as an example.

When I load this file into AIFP, I fill out the initial screen:

When I click continue, this box appears. It lists the aircraft types not listed in the ui_type= lines in my FS installation:

When I click Yes, I get this box displays an empty screen for each missing aircraft type in the drop down box:

I press the Add button and add a relevant AI aircraft from my FS installation. I repeat this for each aircraft type in the drop down box.

After I click Continue, I get this screen. Notice that all the AC#s and reg. nos are 0 or display the aircraft type. THIS is the thing I feel could be changed, since I have specified the aircraft for each aircraft type in the box above. I've given you all the information that AIFP normally gets from the ui_type= line, so I would prefer that AIFP treat that data the same - fill in the AC#s and the reg. no's. with the proper values.

Hope this helps,

 

[gadgets]

Tom, as I explained a while back, the problem is that the aircraft ICAO codes in the timetable do not match the ui-type codes in the aircraft.cfg files. For example, one of the aircraft ICAO codes in the timetable data is "DC-3". The ui_type entry in the aircraft.cfg files you sent me some time ago show "DC-3". Note the included "-". If you change one or the other (or both) so they match, AIFP will find the aircraft - or at least all those that match.

If ui_type is not present in the aircraft.cfg file, AIFP defaults to atc_type in the [general] section and if not there, atc_model in the [general] section.

Don

 

[tgibson]

Ack! Yes Don, I know this. But why can't you go one more step and use the aircraft I add manually, even if the types do not match? If you are not going to do anything with the data you force me to enter manually, then why do you make me select at least one aircraft for each unknown aircraft type, and then promptly ignore all that work? You force me to manually select an aircraft in my FS installation for each unknown aircraft type (a lot of extra work), and then you do nothing with all that information I laboriously entered, other than shove them into the Aircraft List. You could do so much more.

I realize that the aircraft box is currently used to weed out any "bad aircraft" using the Delete button when the user has FS aircraft with matching ui_types, but I fail to see why you cannot use the very same information I enter there manually with the Add button when the ui_types do not match, using the exact same program code? This would result in valid AC# and reg no. values in the flight plans automatically.

Thanks,

 

[tgibson]

PS. When I use a file that has FS aircraft matching the ui_types used in the flight plan file and then add an aircraft using the Add button that has no matching airline code or ui_type to a given aircraft type, that plane is happily included in the flight plans for that aircraft type, using the correct AC# and reg no. All I'm asking is for the same thing when the ui_types do not initially match and I use the Add button in that case.

An example - I used my AIFP Aeronaves plan above with matching ui_types, but then added an SGAA Aero 145 to the DC-3 list of aircraft using the Add button. This plane is included in the AC# and reg no entries of the relevant flight plans. It does not match either the AMX parking code or the DC-3 ui_type. Why aren't mine included when I do the same thing?

 

[gadgets]

Tom, perhaps you should get off your "high horse". I'm not forcing you to do anything. Further, I am able to add/substitute aircraft of differing types and AIFP happily uses them in the flight plans. So, while (at the moment) I have no idea why it (whatever it is) doesn't work for you, you could at least do me the courtesy of specifically reporting a problem before you accuse me of "promptly ignore all that work"

Frankly, given your most recent post, I have no idea what problem you are reporting. You state:

When I use a file that has FS aircraft matching the ui_types used in the flight plan file and then add an aircraft using the Add button that has no matching airline code or ui_type to a given aircraft type, that plane is happily included in the flight plans for that aircraft type, using the correct AC# and reg no.

and add a screenshot to confirm. But then you go on:

Why aren't mine included when I do the same thing?

which seems to contradict your previous post.

Nonetheless, by "reading between the lines" followed by some experimentation, I have discovered that, while an aircraft manually added to the complement of other aircraft of that type automatically "discovered" by AIFP are correctly used in flight plans, if no aircraft of a given type are "discovered" automatically, any you add to that type are ignored. I presume this is the issue of concern. Easily fixed.

Don

 

[gadgets]

Try Development Release 3.2.03(f).

Don

 

[tgibson]

Hi,

Yes, that works, thanks.

 

[boac]

Hi,
I am using AIFP 3.2.03
I haven't complied any flight plans with this version yet.
What I like to do, and it seems no longer can, is to save an aircraft file list on an external drive.
Then, when I am away from FSX I can or could still create flightplans anywhere, and compile them when at my home computer with FSX.

This version now will not let you create a flightplan without FSX on the computer it seems.
A message pops up that states; "No installed Version of Flight Simulator...found. AIFP will now close".

This means I now must be at the computer that has FSX installed.
Can this be changed back, so that flightplans can be created and saved, and complied when at the computer with FSX?
This used to save me bunches of time.
Please consider, and comments are appreciated.
Thank you,
Nick

 

[gadgets]

Nick, that must have been a very early version of AIFP you were using to create FPs independently of FlightSim. Most AIFP functions rely on a version of Flightsim installed on the same computer. It's not something that could easily be undone.

I hope you saved the aifp_.exe file you were using, because I can't provide you another.

The only solution I can think of is for you to install Flightsim on the other computer and then remove most of it (assuming you are concerned about memory usage.)

DOn

 

[boac]

Nick, that must have been a very early version of AIFP you were using to create FPs independently of FlightSim. Most AIFP functions rely on a version of Flightsim installed on the same computer. It's not something that could easily be undone.

I hope you saved the aifp_.exe file you were using, because I can't provide you another.

The only solution I can think of is for you to install Flightsim on the other computer and then remove most of it (assuming you are concerned about memory usage.)

DOn

Luckily I had v3.1.16 saved. So I tried that away from my FSX computer and it seems to load fine.
So I will use that version to create flightplans while away from my computer with FSX, and compile the file when at my computer with FSX.
So I guess I will keep two versions, which is fine.
Thank you for your help!
Nick