Simulation Software Selection

[CristinaLuminea]

I am a masters student and I need to simulate the production process and the energy consumption of a manufacturing company. The aim of this project is to try and find a correlation between the two and lower the energy consumption.

For this I need to choose a simulation software and justify my choice. Unfortunately I do not have much experience in simulation and this is why I am asking for help.

At the moment I am looking at: ProModel, MicroSaint, Witness and Arena. Any information about them would be highly appreciated.

Thank you very much!
Cristina.

[ewilliams]

Will you need to be doing continuous simulation (i.e., the energy consumption) and discrete simulation (the production process) in the same model? ProModel and WITNESS are very good at such "combined" simulation. WITNESS has recently been upgraded significantly in a new version.

[CristinaLuminea]

Thank you ewilliams for replying to my thread. Yes, I would have to do continuous simulation and discrete simulation in the same model.

Also thank you for your suggestions. I already started to review ProModel but didn't get too far yet. Would you have any suggestions of characteristics I would need to look for in order to be sure I am making the right choice?

[ewilliams]

Here is the abstract of a paper published in the 1995 Winter Simulation Conference Proceedings: "Although ProModel for Windows is primarily intended as a discrete event simulator aimed at part manufacturing systems, it supports construction of models that combine continuously varying processes that interact with discrete events. This paper proposes a general approach to modeling of combined discrete-continuous systems, describes the details of a particular example problem, and discusses some modeling issues." The paper presents the ProModel solution to a famous oil-tanker transport problem originally specified by Pritsker [in his textbook Introduction to Simulation and SLAM II]. I say the problem is "famous" because it has traveled far in academia and become a "windmill" against which many software proposals have "tilted." The paper, now 14 years old, is still current in many detailed respects; to my knowledge, the combined continuous+discrete modeling features of ProModel have changed little through many versions. I can send you a photocopy of the paper if you wish to provide a postal address (in those long-ago times, WSC proceedings were bound, not on CD; I have a whole bookshelf full of proceedings 4 feet away from the computer where I am typing this response). WITNESS, similarly, has continuous concepts "fluid," "pipe," "processor," and "tank." One key question to ask yourself is "In my model, will I need logic to express a situation such as 'If a continuously changing variable (e.g. temperature) crosses the value 50 (by going from < 50 to > 50, by going from > 50 to < 50, or in either direction), will I want that threshold-crossing event to trigger a discrete event (e.g., an alarm sounds, a valve closes).'" The answer is most likely yes (based on my experience -- please keep in mind I know barely more than nothing so far about your application). If your answer is "yes," ask yourself how easily the software can do that. In the case of WITNESS, the answer is "easily" -- significantly, the WITNESS help index has an entry "continuous monitoring" which opens up a rich smorgasbord of built-in functions. In the case of ProModel, the answer is "I don't know" for the moment -- I haven't needed to try that lately.

[CristinaLuminea]

Thank you very much for the information ewilliams, I find it really useful. It gives me a starting point. The paper that you are talking about would be good for my research if it would be dated in the last 3-4 years. Unfortunately my mentor would consider it outdated and wouldn't approve me using it as a reference.

Thank you very much again!

Kind regards,
Cristina.