Search the Community

Adam, Gocken, thank you for your input. I have been trying to approach it in a different way so it doesnt create that many entities. However, I couldn't keep a track on the amount of specific goods produced. The model is not entirely finished this is why there might be some empty "decision" steps and worker behaving in a strange way. Any suggestion of how can I keep a track on the goods manufactured and their amount without creating 800k of entities? Let me first explain you the aim. I am working on my final thesis project, where I took a real system under the study. Now I am creating the AS-IS model, which will be the base for further optimization. I want to investigate a potential of implementing cellular manufacturing system. The goal is to see if it's possible to reduce working time, setup times, inventory level and improve on-time delivery. There are 37 workstation where each of them has a stacker machine assigned(for batching of product units into specified by batchsize property groups). Product type depends on the tool(approx. 200 different tools) used in the machine(I want to use a resource to represent a tool however I haven't implemented it yet). Taking into consideration that I'd like to have an overview on the ProductType and its amount produced, as well as specific material utilization for goods manufacturing I have therefore decided to follow this approach. I have been trying to assign a ProductType value into the string state but that was resulting with an error. I want to run simulation for entire year based on the actual orders. Looking forward for any suggestions. Best regards, Dariusz.

It is hard to troubleshoot a model without knowing the system under study and what you are trying to do. That being said, there a couple of things that jump out at me. First, I would recommend testing your logic with a trivial data set, so that you can see what is happening. For example, did you know that you are creating two material source trigger events with the first order? The worker parks itself at output@Source1 on initialization. When the order creation logic starts and a purchase order is released, the worker transfers from the parking station, into the node, and on to free space. That triggers the “ReleaseMaterial” process, which starts the next material release ahead of its scheduled time. Also, the arrivals table is not sorted in date order. That is not a problem for Simio, but it is a problem for your auto-increment logic on the Material Source. For example, the first order actually occurs at 6:01:46 (PO 122137) which is at line 21 in the arrivals table. However, the materials source points that created entity at row 1, so the wrong volume and wrong entity type are created. Again, I would start testing your logic with a very simple data set to see if the model is behaving the way you expect. However, before you go down that route I might suggest re-visiting your conceptual model and model approach. The approach you have started here relies on the creation of more than 800,000,000 entities over the model run . As a sanity check, the default max number of entities (which can be overwritten) is 2500. Creating that many entities (and associated tokens) will require a large amount of computing resources. Even if you verify the model is behaving the way you intend, I can pretty much guarantee you will have performance issues with the current approach. Once you have decided on a tractable approach and have started modeling, there are several built-in tools to help with verification and validation: the trace window and the profile window. Both of these can be used to figure out what is happening in the model—the profile window is particularly helpful for identifying model performance issues. -Adam

there may be many ways for performing this. If your model is not very complex the logic may be: you should track the length of outputbuffer length of source (source@outputbuffer.contents) by using two monitor elements (monitor type, CrossStateChange)...set the InitialThresholdValue for 5 for each one, define two separate processes for each monitor's triggering events; one for above 5 (In this case you should set CrossingDirection positive) and the other for below 5 (and in this case you should set CrossingDirection negative). 1) Create a new process that would be triggered by "above 5 monitör" you should seize worker 2... (source@outputbuffer.contents>5 ---> seize worker 2) 2) And by the same logic Create a new process that would be triggered by "below 5 monitör" you should release worker 2...(source@outputbuffer.contents<5 --->release worker 2...) this is not the end of the story but it is like this.... For example, you should add a storage queue and each time you need to seize worker 2 you have to search storage queue and remove the first entity waiting for a worker (this time worker 2) etc... The second way: In summary, the above logic requires some additional modeling components. Let's try the uploaded file which Works for your purpose with only one process, three steps and a node list. This one is quite easy. But, be sure both will work... and that's it... BottleLine .spfx

Thanks for your help, which was useful. I have now the following approach to dynamically get a WorkerPool using a Search-Step. The first step was to build a new object definition "WorkerPool", which has as its only member a Worker. The second step was then to place six instances of this "WorkerPool" class into the main model and to add all this WorkerPool-instances into a List of objects named "WorkerPoolList". Finally, in the search step, I am searching through the WorkerPoolList for a specific instance of a WorkerPool. As the WorkerPool now is an object and not a Worker, the search step returns only the object.

The easiest way to solve this is to just use two realizations of the same instance of a worker. Set the population of your worker instance to two. The servers will automatically consider the whole population of workers. The other way to do it is to set the servers 'object type' for resources to 'select from list' and give it a list of workers. No process logic required in either instance.

I don't think you can, not in the way you're thinking. However, you could use the search step with a limit of infinity to get separate tokens for every worker of the population. Alternatively, once you are associated with a worker, you can simply use something along the lines of 'Worker.Population....' to get the population functions. You don't even need the search step for this -- if you know the population you are searching, you can directly reference these functions with 'WorkerInstance.Population...'

Try Worker.Population.Capacity.Utilisation, or Worker.Population.Capacity.ScheduledUtilisation.

I have a worker and I want to know utilization of whole population. I would like to have something like Worker.Resourcestate(1) Worker[1].ResourceState(1) gives results for Worker[1] only. Is there any easy way how to do this? Thanks

Hello. Under my M.Sc thesis i´m building a catalog of Simbits to address several characteristics of organizational/information-based business processes/workflows that can be difficult to represent in a simulation model. One of the issues that arose was how to verify the Simbits, i.e, how to prove that the solution really does what i mention it does. For example, one common situation in workflow management systems is the ability of a human resource to redo some work-item (activity/task instance) after its completion. The technical approach in Simio involved a Server object to represent a given task and a Worker object to represent the human resource. The worker has a probability of redo entities at that Server. Therefore, my question is: What is the best approach to show that this solution really works? Using the animation capabilities of Simio with Plots, Pies, Status Labels? Using for example Tally Statistics to record how many times the Worker redone the task during simulation time? Using Experiments maybe? I´m open to read your suggestions. PS: After completing my work i would like to contribute with my Simbits. What i need to do?

Hi, I'm using vehicles in one model and they work in shifts. For example, there are 4 vehicles in shift 1, 8 in shift 2 and 12 in shift 3. However, I'm not sure how I can easily program this, any tips?

Any 'rejected' type tasks are re-evaluated once the worker is available again (they won't just wait indefinately), FYI.

Thanks for the suggestion. It now takes 30 seconds runtime for 1 million ... tokens, I guess that would be. First time I try this low-level approach. I followed your instructions (begin-seize-delay-release-tally). Read the Reference PDF as well. Added a Resource1 though (so there was something to seize?). Can see utilization of Resource1 (seems correct at .8321), but no queueing stats and cannot find any results from that Tally Statistics. Will continue to try. Still, I find it slow. I have a Javasimulation of an M/M/1 that takes only 4 seconds to run 1 million customers (=tokens/entities in Simio). Each customer has to be generated, wait for a resource (seize), delay, release it and ends up in a sink (that tallies). Same computer. The javasim has none of the GUI that Simio has, but then, there's no GUI processing overhead in FastForward/experiment mode. And, the Javasim reports queue and worker statistics (mean, variance). Perhaps Simio is collecting excessive statistics (if so, can it be turned off)? Or is it that I made a mistake by adding a Resource1 (to be seized)?