Grid-based Campaign Generator - Page 3 - ED Forums
 


Notices

View Poll Results: Does GCG interest you?
I'm interested in hearing more about it 144 98.63%
Not interested but good luck 2 1.37%
Voters: 146. You may not vote on this poll

Reply
 
Thread Tools Display Modes
Old 10-26-2015, 03:32 PM   #21
WildBillKelsoe
Veteran
 
WildBillKelsoe's Avatar
 
Join Date: Jun 2009
Location: Cairo, Egypt
Posts: 6,213
Default

Ah I got it. Still be nice to have your plan too. Lots of scenarios your way.

Sent from my SM-T231 using Tapatalk
__________________
AWAITING ED NEW DAMAGE MODEL IMPLEMENTATION FOR WW2 BIRDS

Fat T is above, thin T is below. Long T is faster, Short T is slower. Open triangle is AWACS, closed triangle is your own sensors. Double dash is friendly, Single dash is enemy. Circle is friendly. Strobe is jammer. Strobe to dash is under 35 km. HDD is 7 times range key. Radar to 160 km, IRST to 10 km. Stay low, but never slow.
WildBillKelsoe is offline   Reply With Quote
Old 10-29-2015, 03:18 PM   #22
Haukka81+
Posting Rights Suspended
 
Join Date: Oct 2008
Location: Kuusamo
Posts: 1,356
Default

Quote:
Originally Posted by kontiuka View Post
Thanks for the comments, guys. Appreciated.

Don't have much to report right now. Just doing more testing and of course making sure my program works with the new EDGE engine.

Hoping to give you a substantial status update in early 2016.

kon


Do you need tester? I can test it for you if you like to get help


Sent from my iPhone using Tapatalk
__________________
Oculus CV1 (i5 8400, 16gb ddr4 3000mhz, 1070 8gb OC ) Suomalainen simuryhmä facessa:
https://www.facebook.com/groups/suom...ulaattoristit/

Haukka81 is offline   Reply With Quote
Old 10-29-2015, 04:23 PM   #23
kontiuka
ED Translator
 
kontiuka's Avatar
 
Join Date: Feb 2012
Location: Ottawa, Canada
Posts: 3,387
Default

Quote:
Originally Posted by Haukka81 View Post
Do you need tester? I can test it for you if you like to get help
Thanks, Haukka. I will need some beta (or even alpha) testers so I'll definitely keep you in mind. Things I'll need tested :

- how does GCG install on different systems
- how does it run on different systems (are there crashes?)
- how easy is it to use (is it intuitive? is the user guide descriptive enough?)
- are the generated missions buggy or illogical (ex. 1 tank attacking 40)

I'll likely be asking for help sometime next year. Don't know exactly when though.

kon
kontiuka is offline   Reply With Quote
Old 10-29-2015, 06:48 PM   #24
Haukka81+
Posting Rights Suspended
 
Join Date: Oct 2008
Location: Kuusamo
Posts: 1,356
Default

Quote:
Originally Posted by kontiuka View Post
Thanks, Haukka. I will need some beta (or even alpha) testers so I'll definitely keep you in mind. Things I'll need tested :

- how does GCG install on different systems
- how does it run on different systems (are there crashes?)
- how easy is it to use (is it intuitive? is the user guide descriptive enough?)
- are the generated missions buggy or illogical (ex. 1 tank attacking 40)

I'll likely be asking for help sometime next year. Don't know exactly when though.

kon

Im always ready to help other flight sim guys

Im nurse in real life so i know how to make nice reports.

And tacview will help too.

But just pm me when you need help, i will make my best to get this thing forward !

-haukka81




Sent from my iPhone using Tapatalk
__________________
Oculus CV1 (i5 8400, 16gb ddr4 3000mhz, 1070 8gb OC ) Suomalainen simuryhmä facessa:
https://www.facebook.com/groups/suom...ulaattoristit/

Haukka81 is offline   Reply With Quote
Old 10-30-2015, 09:21 PM   #25
AndyHill
Member
 
Join Date: Dec 2004
Location: Finland
Posts: 236
Default

I've been involved in testing and even kind of in the design efforts of some 3rd party dynamic engines for various games so I might as well start following this one as well. I'm not sure I completely understand your design, but it might be quite close to what I consider a good model for a dynamic campaign engine; the Il-2 DCG by Lowengrin (http://www.lowengrin.com/news.php).

DCG is a deceptively simple (and this is a good thing) design that has immense room for growth. Basically it consists of a table of units for both sides and a (hand made) grid (is this similar to your concept of grid?) of nodes and routes that are used by the units to move and fight. The units move towards each others' spawns points along a network (grid?) of nodes and routes, the units are mostly trains, supply columns, infantry columns and tanks. When they meet on the battlefield they fight it out and when tanks occupy a node, their presence reduces the number of defending infantry every "turn" until they conquer it (or get killed) at which point the support units roll in and fill the node with their own infantry (actually they usually roll right into the combat zone anyway all the time).

The node-route -networks are mostly simple lines with occasional branches and when a unit reaches a branching point it basically randomizes the direction it takes. There are maybe about two dozen (hand placed) nodes per campaign and the distance between them is usually several to a few dozen kilometers. When a unit (such as a tank platoon) is destroyed it goes back into a pool and before each mission there's a chance it will respawn in either of the end nodes.

That's basically it, sounds simple, is simple and has produced tens of thousands of man hours worth of fun for our smallish group of people during the last decade. You don't need a super complicated system to entertain people, that's the most important lesson learned from DCG. A dynamic campaign generator is probably VERY MUCH WORTH DOING even if it's not the most perfect and complicated thing ever.

Most importantly DCG would have so much potential for growth. Resource pools for units would probably make the campaigns flow better (currently they are highly stabilizing designs due to the winning side's supply routes getting massively longer and eventually the defending side gets to fly their ground attack planes to target before the other side has a chance to cover them) and some kind of rudimentary AI and a wider network of nodes and routes would make the campaigns much more organic. And a million other things.

The improvements are unfortunately unlikely to happen, since Lowengrin himself is quite understandably tired and probably busy with other stuff - and the project isn't open source so no-one else can pick it up and continue. I'm not a very capable programmer and my ideas probably aren't that brilliant, but if you go open source I would at least try to help in any way I can. If you're for example trying to monetize the campaign that's of course fine and ED's business model even supports such a thing nowadays. In that case if the campaign is good enough I can probably at least review it for a games magazine with at least a few readers potentially interested in such a thing. However, if that's not your intention, I highly recommend open sourcing the effort at some point.

If you're interested (and how could you not be interested in the pearls of my infinite wisdom), here's how I'd subdivide a dynamic campaign:

- The strategic stage that defines the goals and available units (in the beginning this could be a list of units that spawn at the starting point after getting killed like Il-s DCG)

- Tactical stage that define where the units go on the map and stuff like that (at first this could be a simple linear maximum speed progression towards enemy spawn with randomness at possible crossroads like in DCG)

- Mission generation stage that generates the actual mission file for a specific simulator and for example defines which units actually appear on the map and which planes are flyable etc. (in the beginning it can simply be every unit in every mission)

- Mission analysis and simulation stage is where the results are analyzed and if necessary simulated (for example if the mission duration is thought to be one hour and the player quits at five minutes after a botched take-off, this also makes it possible to have bigger campaigns where your role is realistic-sized and not a hero who has the power to kill all existing units in one mission. Of course in the beginning this isn't necessary)

I think that might be pretty much it. Every single stage can be incredibly simple and still provide a fantastically entertaining campaign (which is true for almost any design, never forget that!) and endless room for improvement and expansion.

Anyway, I wish you the very best of luck with your project, DCS needs dynamic campaigns more than anything else at the moment. I don't know if I can actually be of any help, but I'll do what I can (and sparing you from my wonderful ideas and opinions probably belongs in the I can't -category, sorry) if you need anything.

PS. How about multiplayer? That's of course a huge thing for me and especially my (not imaginary, I swear!) friends and it might not be trivial in DCS.
__________________
My blog full of incoherent ramblings on random subjects: https://anttiilomaki.wordpress.com/
AndyHill is offline   Reply With Quote
Old 10-31-2015, 11:28 AM   #26
Ripcord
Member
 
Ripcord's Avatar
 
Join Date: Mar 2010
Location: Houston TX
Posts: 657
Default

What this sim truly lacks, and would be not that hard to implement in my uneducated opinion, are campaign variables. Blow up this, kill all units in that zone, etc., a flag is triggered. We have that in the mission editor now and have had it for years. But a flag tripped on the first mission cannot be read by the following mission. So we need a series of flags that CAN be read by the next mission in a series.

Janes FA18 had those, and they were called campaign variables. Campaign variables could be set in a way to alter the flow of the missions in a campaign, particularly if some global event happened. Now all that was very limited, in that sim, but it would make a TON of sense having this set of persistent variables or flags in THIS SIM. If you had enough of those, you could use those to program what units were killed and where the burning wrecks could be placed for the following mission.

What we do have, however, is the campaign ladder logic itself. You have the ability to generate a score from zero to 100. So a really bright person could harness this and create his own logic system that scores a mission in such a way that it not only tracks the outcome from a win/loss perspective but it will more or less designate a list of survivors from the mission, and you can have quite a number of those -- remember you can have up to 100 different outcomes, not just win or lose. You could, if you really wanted, have 100 different possible missions at each rung in the campaign ladder and in doing so you could really simulate a moving FLOT. It would require having a LOT of missions in the campaign admittedly, but it would work.

When I read Kon's description here this is immediately what I thought of.
__________________
<a href=http://forums.eagle.ru/signaturepics/sigpic65507_1.gif target=_blank>http://forums.eagle.ru/signaturepics/sigpic65507_1.gif</a>

Last edited by Ripcord; 10-31-2015 at 11:32 AM.
Ripcord is offline   Reply With Quote
Old 11-02-2015, 02:37 AM   #27
kontiuka
ED Translator
 
kontiuka's Avatar
 
Join Date: Feb 2012
Location: Ottawa, Canada
Posts: 3,387
Default

Quote:
Originally Posted by AndyHill View Post
I've been involved in ...
Hi, Andy.

Thanks for the description and the link. Sounds like an intriguing project. Ya, I would say there are some similarities between DCG and my campaign generator - most obviously, the use of grid points. I think the two strategy engines are probably quite different though.

If I were to describe how mine worked, I would say picture a general or group of generals in a room leaning over a map and assessing several possible troop movements. For each scenario, they predict how the enemy is likely to react, how many losses will be sustained, and how much ground will be covered. Then the generals pick the most favourable scenario. This is just an analogy and very high level but hopefully gives you an idea of what I'm trying to achieve.

No, I don't plan on monetizing. Just wanted to do something for the community. I'll always leave my options open though.

As for multi-player. Well, I play single player primarily and that will be my primary focus in my design but, yes, there will certainly be the flexibility for a single player mission to be made multi-player. At least, to a limited degree.
kontiuka is offline   Reply With Quote
Old 11-02-2015, 03:38 AM   #28
Sierra99
Campaign Testers
 
Sierra99's Avatar
 
Join Date: Apr 2012
Location: Orlando Florida
Posts: 1,722
Default

So if I understand you correctly, you initially place both side in a "starting" position, play the mission and quit and the GCG develops the follow on mission?
__________________

Gigabyte GA-Z68XP-UD3, i7-3770K CPU @ 3.70GHz, 32GB G.SKILL Ripjaws DDR3 @ 1600Mhz, ZOTAC GeForce® GTX 1070 Ti AMP Extreme , Windows 10, Thrustmaster Warthog, Cougar MFDs x3, Saitek Rudder pedals and TrackIR 5.
Sierra99 is offline   Reply With Quote
Old 11-02-2015, 03:01 PM   #29
kontiuka
ED Translator
 
kontiuka's Avatar
 
Join Date: Feb 2012
Location: Ottawa, Canada
Posts: 3,387
Default

Quote:
Originally Posted by Ripcord View Post
What this sim truly lacks ...
Hi.
The way I’m doing things, I won’t really need to use flags. When a mission is played, any units killed are written to a log file. When the user generates the next mission, the first thing GCG does is remove all those dead units from the active units list (I’ll eventually be implementing reinforcement logic but that will come later). Then, like I described in my last post, GCG will do an analysis of the entire battlefield with the revised active units list and generate the next moves. And, yes, the wrecks of the dead units will be placed in subsequent missions.
Because the missions are built dynamically, there won’t be a need to hand-craft a bunch of missions for branching purposes. Designing a bunch of missions would be a tremendous amount of work which is why I decided to work on GCG in the first place.

Last edited by kontiuka; 11-02-2015 at 03:29 PM.
kontiuka is offline   Reply With Quote
Old 11-02-2015, 03:40 PM   #30
kontiuka
ED Translator
 
kontiuka's Avatar
 
Join Date: Feb 2012
Location: Ottawa, Canada
Posts: 3,387
Default

Quote:
Originally Posted by Sierra99 View Post
So if I understand you correctly, you initially place both side in a "starting" position, play the mission and quit and the GCG develops the follow on mission?
Yes, that is exactly right. My hope is that someone with even just a basic knowledge of the DCS mission editor will be able to design a "starting" mission using a GCG template mission and GCG will take care of all mission generation from that point on. I'll likely include a few pre-made starting missions from fairly simple to fairly complex so people can try things out without having to do any mission design.
kontiuka is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

All times are GMT. The time now is 11:42 AM. vBulletin Skin by ForumMonkeys. Powered by vBulletin®.
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.