I have a single-lane road expanding into a multi-lane road with a stop line at the end of it. As the car moves from the single-lane section to the multi-lane section it selects the lane seemingly randomly. However, this causes the multi-lane section to back up much quicker than what would happen in real life as one lane will fill up while the other is still empty.
Is there a way to control this behaviour?
I use the following approach in this type of situations:
define a function called fnSelectLane(car)
manually assign the lane by conditions
You can also check how many cars are in each lane with this function int nCarsOnLane(boolean isOnForwardSide, int laneIndex).
Long story short, you can manually assign the lanes, if you don't like the default setup.
Related
This model simulates cars pulling up to pumps at a gas/petrol station. I have a custom car agent 'MyCar' that passes through an upstream stopline. At this stopline it calls the agent's function which checks the availability of the two pump stations in the roadNetwork in Main. This function returns a RoadNetworkConstant 'PASS_THROUGH_STOP_LINE' or 'STOP_BEFORE_STOP_LINE'.
I'm unable to pass this constant value back into the 'Stop line behavior' property in a CarMoveTo block (for a different stop line further downstream but before the car reaches the pumps) in Main even when I start typing the name of my agent 'MyCar'.
Is it possible to change the Stop Line Behavior dynamically? If so, is this a resonable approach? If not, what would you suggest?
You need to use 'car', not 'MyCar'. Check out the magic light bulb and how keywords work for you: https://www.google.com/url?q=https://www.benjamin-schumann.com/blog/2016/2/4/the-magic-lightbulb-and-how-it-can-help-your-anylogic-modelling&sa=U&ved=2ahUKEwiuz-ytq9D8AhWsSvEDHZccDTUQFnoECAkQAg&usg=AOvVaw3XapUwKyi_Bi9_2JjGrN0b
I need to create a model of a simple train station where I want to have the "unwanted" option to have trains collide without the model crashing with one exception. Is there any chance to avoid the model crashing maybe stopping the two trains colliding when it happens (or just a moment before)?
Thanks a lot!
Drag and drop Rail Settings and it has a few things for conflict resolution.
On car hit car section allows you to type in any code.
Here are the local variables that you can use:
You can also watch this video from AnyLogic: https://www.youtube.com/watch?v=ERvdz0w_Zgk
You can also simply use the outHit port on the TrainMoveTo block:
Best do the simple rail-model step-by-step tutorial (hump yard) as it shows how to use it
I want to implement a connected autonomous vehicle fleet in AnyLogic. Because of this, a vehicle agent has the ability to change its route. My understanding is that using the standard "MoveTo" block lets you move the vehicle agent only once by telling it what road to move the agent to.
For example:
Any vehicle following the above block can be created on any road that is specified in road4Source, then move to another road specified in moveToRoad2 block and then will be destroyed/deleted/discarded when it reaches the endOfRoute block.
However, since the vehicle can choose the route for itself, I only know the Origin and the Destination of the vehicle and not sure what route it will take. So, the number of MoveTo blocks should ideally be dynamic.
Considering this, is it possible to reuse MoveTo blocks (consider it a kind of a while loop) so that that the vehicle remains in the MoveTo block until it reaches the Destination (a particular road in my case)?
Something like this:
So that in the Road argument of moveToRoad2, I can put something like(could be wrong just for illustration):
car.getRoad() == car.destination ? endOfRoute : car.getRoad()
where car.getRoad() would return the road that the vehicle wants to go to and if it is equal to the destination of the car then it would go the endOfRoute block.
EDIT 1
I understand now that if there are two connections to the same port, AnyLogic will throw an error:
Out port can't be connected with more than one In Port.
So is there any other way to achieve what I want to do here?
First, this is the configuration you want:
Then you need to generate a collection of roads that the car will take... This collection is generated in the carSource... or you can update it as you progress in the "loop".
the car agent must have a variable that will start with 0 and will increment +1 every time it exits carMoveTo...
So the new road that the car will take is:
collection.get(car.variable)
And on exit of carMoveTo:
car.variable++;
In the selectOutput the condition for true would be:
agent.variable==collection.size()
I hope this helps...
Nevertheless, if you have too much traffic, you will see big problems in your model because the car doesn't necessarily recognize traffic after the destination in the carMoveTo so you may have crushes... or maybe not... depends on different factors.
I am having trouble with my Unreal Engine 4 project. I'm very new to this and i don't really understand what's going on. I have made ladder functionality so the character can climb up the ladder and stand on the Static Mesh that is on top. But when i want to go down the ladder i want to trigger the "Allow Down" function on Actor "BP_Ladder" but the cast is failing everytime. what is causing the casts to fail?
I've looked around and other people with cast failed problems have mostly had the names wrong but my ladder is called "BP_Ladder" and that is what i'm casting to so that leaves me really confused.
My BP_Dude blueprint (this is being called every tick)
My BP_Ladder blueprint (this is what i'm trying to trigger)
The goal for this function is that the collision of the static mesh will turn off and then allow me to move down the ladder like normal.
I'd really appreciate your help with this, its my first couple of days using unreal engine and the Epic Games tutorials i followed didn't show all of the blueprints so everyone was left helpless with half broken blueprints.
Understanding Casts
In its simplest form, if your Actor can be cast to BP_Ladder, then it will return the casted object on the output pin.
Just to ensure you know how a cast works, I have to point out that you can't cast a given type to some unrelated arbitrary type; it has to be "castable" to the cast destination. So if you think that the Actor object returned by your overlap is genuinely a BP_Ladder, or a blueprint class that derives from BP_Ladder, then you should be OK. But that has to be the case; otherwise it will fail every time.
https://docs.unrealengine.com/en-us/Engine/Blueprints/UserGuide/CastNodes
Sorry if I'm being patronising, but if a cast isn't working 9/10 it hasn't been used correctly.
Debugging
OK that out of the way, if you think you are genuinely casting to the correct type and it's still failing, then you'll need to debug your blueprint with the objective of finding what type is being given to the cast node which results in the failure.
Select the cast object in your blueprint.
Press F9 to create a breakpoint; a red circle should appear on the top left of your cast box
Run your game in PIE mode (Combo next to play, New Editor Window (PIE))
Put the game in a scenario where your cast will fire (I presume touch the ladder)
The breakpoint should trigger; your game window will go grey and you'll get a large red arrow pointing down towards your cast box where you placed the breakpoint
Hover over the Object pin on the input side of the cast box. It should show a alt-text containing details of the variable.
Inside the alt-text box, look for Current value; this should show you the current object type that you are about to cast. You need to ensure that this value is what you expect.
https://docs.unrealengine.com/en-US/Engine/Blueprints/UserGuide/Debugging
Example
I've taken a screenshot of a working game; in this you will see:
a Breakpoint (the red circle) is shown on the function node (the box)
the Execution node (red arrow) is Create Rolling Stock
the Current value of the variable is DepotComponent...Depot
Final thoughts
There's a couple of gotchas when using blueprints; one of them in this case could be that you can have two classes with the same name (although they might have different internal "script names" - effectively a fully qualified path). When you debugging an object variable, make sure you match the entire path to the location of your asset in your content folder; this will ensure that you are indeed attempting to cast to the object you think you really are.
Another classic gotcha is "hot reloads". Sometimes the editor needs to reload a module after an on-the-fly build but a class conflict occurs internally; this results in the new version of the class getting a different name (prefixed with HOTRELOADED). Watch out for this; it can cause you to be dealing with two distinct types when casting and you don't even realise. A real sanity-sapper. If in doubt, close and re-open the editor; it fixes it every time.
On Begin Overlap is only triggered when you START an overlap, so if you have some logic that sets your "can climb" to false you will need to move outside of the collider and back into it again for another Begin Overlap event.
My goal is to build a 5x5 grid of images. In the following code, row, col and rowcol were created as variables local to the sprite, and newcol, newrow and cats are global. (By the way, is it possible to tell which variables are local and which are global? It's easy to forget or make mistakes.)
The result is a 5x1 grid only, as seen here.
I am unclear as to the order of execution of these statements. Does when I start as a clone get called before or after add_cat gets called the second time? My tentative conclusion is that it gets called afterwards, yet the clone's global variables seem to contain their values from beforehand instead.
When I attempted to debug it with ask and say and wait commands, the results varied wildly. Adding such pauses in some places fixed the problem completely, resulting in a 5x5 grid. In other places, they caused a 1x5 grid.
The main question is: How to fix this so that it produces a 5x5 grid?
Explanation
Unfortunately, the execution order in Scratch is a little bizarre. Whenever you edit a script (by adding or removing blocks, editing inputs, or dragging the entire script to a new location in the editor), it gets placed at the bottom of the list (so it runs last).
A good way to test this out is to create a blank project with the following scripts:
When you click the green flag, the sprite will either say "script one" or "script two", depending on which runs first. Try clicking and dragging one of the when green flag clicked blocks. The next time you click the green flag, the sprite will say whichever message corresponds to the script you just dragged.
This crazy order can make execution incredibly unpredictable, especially when using clones.
The solution
The only real solution is to write code that has a definite execution order built-in (rather than relying on the whims of the editor). For simpler scripts, this generally means utilizing the broadcast and wait block to run particular events in the necessary order.
For your specific project, I see two main solutions:
Procedural Solution
This is the most straightforward script, and it's probably what I would choose to go with:
(row and col are both sprite-only variables)
Because clones inherit all sprite-only variable values when they are created, each clone will be guaranteed to have the correct row and col when it is created.
Recursive Solution
This solution is a bit harder to understand than the first, so I would probably avoid it unless you're just looking for the novelty: