Anylogic: This should be simple but I just cannot find it in the help files..
On creating a new agent instance, we know there are four parameters and what they are, but not in what order they were defined. Lets say parameters are "type_of_car" (String), "number_of_pax" (Integer), "automatic" (boolean), "fuel_capacity" (double). Now when calling: new myagenttype("ford", 5, false, 55) the agent gets created as a ford with 55 pax, manual and 5 liter fuel capacity - which is all wrong. (it seems the definition order of the parameters in the agent definition are in a different order)
How do we include the parameter name (or definition) when we call new agenttype() to avoid this problem, ensuring the right value gets assigned to the right parameter?
The problem originates because of a bug in Anylogic's logic in triggering functions. We have a Split which creates a new agent and assigns the agent properties on the "On exit copy" - event, however what Anylogic does is it creates the agent, forward it to the next logical block (a decision node), then execute the code of the decision - all wrong now for the agent's properties are undefined - and only then executes the "On exit copy" event which assigns the agent's properties. Very frustrating.
This is actually not a bug, this is defined in the simulation experiment properties, in the randomness section, with the "selection mode for simultaneous events" property. The default is LIFO, but if you want the opposite behavior in your case you should use FIFO... I always use LIFO too, and in these cases, sometimes I might use a 1milisecond auxiliary delay between the split and the next block in order to control the order manually... if you do that, you will solver your problem in fact... just use a 0.001 miliseconds delay after your split
Now from your Agent Type problem, the arguments for your class constructor should be written in order, and the only way to know the correct order is by using the autocomplete feature when you write new agenttype()... the autocomplete will tell you the order in which you should write your constructor arguments.
Related
In my model I am transporting (path guided) several types of products (motors) by an agv trough a manufacturing line with 27 cycles. It´s a flowing manufacturing line. That means the product gets manufactured while the agv is constantly running.
To model that I created an agent population called "motors" with parameter "axelType" (string) which is loaded from column "axel_type" in database "manufacturing_sequence" (local excelsheet) and is placed on Main.
Each motor is placed on an transporter "agvAssembly" (in Flowchart as: Transporter) and runs from node "locationCycle1" all the way to "locationCycle27".
Now I want to change the transporters speed at each of the 27 cycle nodes dependent on the currently loaded motor. To do that I got another database called "speeds_axel" which includes all the needed speeds for the cycles and respective parameter name for axelType (column axel_type).
So now, when the transporter enters a node I have to check first the nodes name. Than I want to read out the parameter "axelType" of the currently in that node entered agent "Motor" and search in the database for the respective speed.
In the block "transporter Fleet" - "On enter node:" I wrote as an example for cycle 1 the following:
if (node == locationCycle1) {
unit.setMaximumSpeed(
selectFrom(speeds_axel)
.where(speeds_axel.axel_type.eq(motor.axelType))
.firstResult(false, speeds_axel.cycle1)/60.0,MPS);
}
When running the model I get the following error:
"motor cannot be resolved to a variable"
(location: TransporterFleet)
I think that error accures because my approach doesn't specify which motor I mean. How can I clarify to AnyLogic that I always mean the current motor which enters the node?
I need something like this:
if (node == locationCycle1) {
unit.setMaximumSpeed(
selectFrom(speeds_axel)
.where(speeds_axel.axel_type.eq("get motor which is currently in locationCycle1".axelType))
.firstResult(false, speeds_axel.cycle1)/60.0,MPS);
}
the biggest problem in your model is that you don't follow conventions. An agent type should be named with an uppercase first letter
Why is this important? Because you want to make the difference between Motor, motor, and motors
Motor is the class (or Agent Type)
motor is an instance of that class (or agent type)
motors is a population.
Since you don't follow this convention, you make mistakes of this kind since motor is a class
in your case when you do motor.cycle1
if you followed the conventions, you would be doing
Motor.cycle1 (which is obvioulsy wrong)
Note that Motor is the Agent Type name, and what you really want is to know for a particular motor what the value of cycle1 is
The first thing you need to do with this model, is get back to using the conventions, and this will probably solve this problem and many problems in the future.
Just because you have a population called motors (which the agents within your flow 'come from' --- though it looks like they don't actually; see later), this does not mean that you can 'magically' refer to the current agent via the singular form motor at any point in the code. (And, similarly, your agent type being Motor does not mean you can use the lower-case form to refer to 'the current one'.)
In Transporter Fleet block "On enter node" actions, you refer to the current transporter via the unit keyword; see the help page. The motor is not the transporter; it's the thing being transported.
So, just use the "On exit" (or, better, "On at exit") actions of your MoveByTransporter blocks (which trigger on arrival at the destination node), where you can refer to the current motor (the agent in the flow) via the agent keyword; again, see the block help.
But, from a quick glance, there appear to be a few important changes/simplifications you should also make (though hard to definitively tell without details of all your code):
You are using a Source block to add agents (motors) to the flow when you have already created the motors in your population. You should be using an Enter block instead to add the (already-existing) Motor agents to the flow at the appropriate time (triggered by code: I imagine you want one to start at model start time, requiring code in Main's "On startup" action, and then others added either at certain simulation times or when earlier motors reach a certain stage in the process...). If your population was actually just for 'templates' of data for each motor axle type (and you would then potentially create multiple instances of motors for a given axle type) then your current logic might make sense, but your screenshot of your manufacturing_sequence table makes clear that isn't the case. (If you did go that route, you would want to rename your population so that its purpose was clear.)
It looks like you have a 'looping' process (with the same process behaviour for all your in-sequence nodes), so you should look to make your process generic with an explicit loop (so you don't have near-copies of blocks for every node in your sequence). There is a fair amount of detail in terms of how you do this but roughly:
Track the node the motor is currently in (or its sequence number) via a variable in the Motor agent.
Use a collection (List) which contains the nodes-in-order to determine where it has to go to next.
Your process would have a MoveByTransporter --> Delay --> SelectOutput sequence (plus TimeMeasureStart/End if you're using them), where SelectOutput loops back to the MoveByTransporter if the motor is not yet at the last node.
I want my agent to be of a single type. Where type contains a list of 3 options. And the allotment should be based on probability. E.g. let's say I want to allot 1: 30%, 2: 50% and 3: 20% to each of the agents generated in the source of my main tab.
I tried with one of the ways by declaring th parameter as int and then writing randomTrue(0.3)?1:randomTrue(0.7)?2:3 in the default value. But every time the agent comes with the same value of 2.
Please can anyone help me with this??
And if I try to allot the parameters in the main window at any of the blocks, do they get attached to the specific agent that passes through that block or its value just gets updated for that time until the other agent passes? Actually I have to check each and every agent for the parameter and then send it through a specific output path from the selectoutput block.
Your code does not actually change anything in the created agent.
Create a parameter `myType´ of type Integer.
In your source code write:
agent.myType = randomTrue(0.3) ? 0 : randomTrue(0.7) ? 1 : 2
Make sure your Source block actually creates agents of the Agent type that holds the myType parameter.
Study some of the example models and tutorials, it is covered in many places :-)
For a project I'm trying to select the right output if a parameter of the created agent is true. The agent is created in the main but goes into a different agent for the selection of the output. If I use a select in the main and use the statement for the parameter as condition it works fine. Whenever it enters the agent and I use the same statement it gives an error (Description: product1 cannot be resolved or is not a field. Location: testCase5/Transport/selectOutput - SelectOutput).
I tried using different conditions like main.agent.product1 . Can't seem to solve the issue. I know that it has to do something with declaration probably but I'm clueless at the moment.
//for the condition that works if I put the selectOutput in the main
agent.product1
//for the condition in the agent I tried the same but didn't work. also tried
main.agent.product1 // Gives the same error but for (Description: agentcannot be resolved or is not a field.)
(Main)agent.product1
//And a few more
In AnyLogic each block of a project flow (regardless if from the process modeling library, the pedestrian library or another) has the type of Agent that is flowing through it defined in its properties. You can find this setting as Agent type in the block's properties under Advanced.
If you fail to set this correctly (meaning this setting equals the type of Agents you let flow through it), it will still flow through, BUT: you cannot access the specific properties of the Agent type.
That said, usually you do not even have to think about, because AnyLogic has an automatism there: The first block of your process flow, typically a Source-block or a Enter-block is the only block where you have to make the correct setting, all attached blocks get it then automatically!
However this does not work when:
You have separate flows (eg. you leave one flow through an Exit-block and enter another by an Enter). You'll manually have to put the correct setting also for the first block of the second process flow.
Your process flow is continued inside of another Agent/Custom Block, as in your example. AnyLogic doesnt realize the process flow inside your custom block is logically connected to the outside process flow in main. You have to specifiy manually again the Agent type flowing through in the first block of your process inside your Custom Block.
the basic idea behind the modeling issue is a breakdown of a production machine.
I would like to model this by setting the arrival rate (simply arrivals per second) to zero (Source.rate = 0). After the machine is repaired, the arrival rate is set to its actual value again (e.g., Source.rate = 5). While the first command does the job, the second does not seem to have any effect, i.e. new agents are not created.
The segment of the model is rather simple: Source --> Select Output (decision about breakdown) --> true: go on in production; false: delay (repair machine) --> go on in production.
Source.rate = 0 is called at the out port (false) of "breakdown" and Source.rate = 5 at the out port of "repair".
https://i.stack.imgur.com/hqGoI.png
Of cause, this issue might be modeled differently (e.g., using hold with disabled "forced pushing"), however, it is not clear for me why my approach does not work.
Thanks in advance!
Instead of using source.rate=5; use source.set_rate(5);
To expand on Felipe's answer with an explanation:
Instead of using source.rate=5; use source.set_rate(5);
rate is effectively a Parameter (in the AnyLogic sense) of the Source block. (All AnyLogic's Process Modeling blocks are actually themselves Agents developed by AnyLogic, and thus with Parameters, Variables, etc.)
You can set an AnyLogic Parameter directly (via just assigning a value as you did), but they also all have a set_<parameter name> method (function) which should really always be used instead because this triggers any internal on-change logic for this Parameter. It is only this triggered logic (internal to the Source block) which causes the Source to 're-evaluate' the rate properly.
(You can use on-change logic for Parameters in your own models, and need to do so when altering a parameter requires some 'adjustments' to the rest of the model; i.e., in situations where the change doesn't 'just work' due to other bits of the model reading the new value after the change point.)
I don't know why your model doesn't work (maybe more details of your model is needed), but a simple solution which I tested and worked, is as below:
You can set the source's "Type of arrival" to "calls of inject() function", add an event to your model and set its "Trigger type" to "Rate" and set its rate value to 5. Then in action code of the event use below code:
if(yourCondition)
{
source.inject(1);
}
I hope it helps you.
We are developing an ABM under AnyLogic 7 and are at the point where we want to make multiple simulations from a single experiment. Different parameters are to be set for each simulation run so as to generate results for a small suite of standard scenarios.
We have an experiment that auto-starts without the need to press the "Run". Subsequent pressing of the Run does increment the experiment counter and reruns the model.
What we'd like is a way to have the auto-run, or single press of Run, launch a loop of simulations. Within that loop would be the programmatic adjustment of the variables linked to passed parameters.
EDIT- One wrinkle is that some parameters are strings. The Optimization or Parameter Variation experiments don't lend themselves to enumerating a set of strings to be be used across a set of simulation runs. You can set a string per parameter for all the simulation runs within one experiment.
We've used the help sample for "Running a Model from Outside Without Presentation Window", to add the auto-run capability to the initial experiment setup block of code. A method to wait for Run 0 to complete, then dispatch Run 1, 2, etc, is needed.
Pointers to tutorial models with such features, or to a snip of code for the experiment's java blocks are much appreciated.
maybe I don't understand your need but this certainly sounds like you'd want to use a "Parameter Variation" experiment. You can specify which parameters should be varied in which steps and running the experiment automatically starts as many simulation runs as needed, all without animation.
hope that helps
As you, I was confronted to this problem. My aim was to use parameter variation with a model and variation were on non numeric data, and I knew the number of runs to start.
Then i succeed in this task with the help of Custom Variation.
Firstly I build an experiment typed as 'multiple run', create my GUI (user was able to select the string values used in each run.
Then, I create a new java class which inherit from the previous 'multiple run' experiment,
In this class (called MyMultipleRunClass) was present:
- overload of the getMaximumIterations method from default experiment to provide to default anylogic callback the correct number of iteration, and idnex was also used to retrieve my parameter value from array,
- implementation of the static method start,
public static void start() {
prepareBeforeExperimentStart_xjal( MyMultipleRunClass.class);
MyMultipleRunClass ex = new MyMultipleRunClass();
ex.setCommandLuneArguments_xjal(null);
ex.setup(null);
}
Then the experiment to run is the 'empty' customExperiment, which automatically start the other Multiple run experiment thru the presented subclass.
Maybe it exists shortest path, but from my point of view anylogic is correctly used (no trick with non exposed interface) and it works as expected.