Introduction

The topic is modeling of the operation of (manufacturing) systems, e.g. semiconductor factories, assembly and packaging lines, car manufacturing plants, steel foundries, metal processing shops, beer breweries, health care systems, warehouses, order-picking systems. For a proper functioning of these systems, these systems are controlled by operators and electronic devices, e.g. computers.

During the design process, engineers make use of (analytical) mathematical models, e.g. algebra and probability theory, to get answers about the operation of the system. For complex systems, (numerical) mathematical models are used, and computers perform simulation experiments, to analyze the operation of the system. Simulation studies give answers to questions like:

The operation of a system can be described, e.g. in terms of or operating processes.

An example of a system with parallel operating processes is a manufacturing line, with a number of manufacturing machines, where product-items go from machine to machine. A surgery room in a hospital is a system where patients are treated by teams using medical equipment and sterile materials. A biological system can be described by a number of parallel processes, where, e.g. processes transform sugars into water and carbon-dioxide producing energy. In all these examples, processes operate in parallel to complete a task, and to achieve a goal. Concurrency is the dominant aspect in these type of systems, and as a consequence this holds too for their models.

The operating behavior of parallel processes can be described by different formalisms, e.g. automata, Petri-nets or parallel processes. This text uses the programming language Chi, which is an instance of a parallel processes formalism.

A system is abstracted into a model, with cooperating processes, where processes are connected to each other via channels. The channels are used for exchanging material and information. Models of the above mentioned examples consist of a number of concurrent processes connected by channels, denoting the flow of products, patients or personnel.

In Chi, communication takes place in a synchronous manner. This means that communication between a sending process, and a receiving process takes place only when both processes are able to communicate. Processes and channels can dynamically be altered. To model times, like inter-arrival times and server processing times, the language has a notation of time.

The rationale behind the language is that models for the analysis of a system should be

Verification of the models to investigate the properties of the model should be relatively effortless. (A model has to preserve some properties of the real system otherwise results from the simulation study have no relation with the system being modeled. The language must allow this verification to take place in a simple manner.)

Experiments should be performed in an straightforward manner. (Minimizing the effort in doing simulation studies, in particular for large systems, makes the language useful.)

Finally, the used models should be usable for the supervisory (logic) control of the systems (simulation studies often provide answers on how to control a system in a better way, these answers should also work for the modeled system).

Chi in a nutshell

During the past decades, the ancestors of Chi have been used with success, for the analysis of a variety of (industrial) systems. Based on this experience, the language Chi has been completely redesigned, keeping the strong points of the previous versions, while making it more powerful for advanced users, and easier to access for non-experts.

Its features are:

  • A system (and its control) is modeled as a collection of parallel running processes, communicating with each other using channels.

  • Processes do not share data with other processes and channels are synchronous (sending and receiving is always done together at the same time), making reasoning about process behavior easier.

  • Processes and channels are dynamic, new processes can be created as needed, and communication channels can be created or rerouted.

  • Variables can have elementary values such as boolean, integer or real numbers, to high level structured collections of data like lists, sets and dictionaries to model the data of the system. If desired, processes and channels can also be part of that data.

  • A small generic set of statements to describe algorithms, assignment, if, while, and for statements. This set is relatively easy to explain to non-experts, allowing them to understand the model, and participate in the discussions.

  • Tutorials and manuals demonstrate use of the language for effective modeling of system processes. More detailed modeling of the processes, or custom tailoring them to the real situation, has no inherent limits.

  • Time and (quasi-) random number generation distributions are available for modeling behavior of the system in time.

  • Likewise, measurements to derive performance indicators of the modeled system are integrated in the model. Tutorials and manuals show basic use. The integration allows for custom solutions to obtain the needed data in the wanted form.

  • Input and output facilities from and to the file system exists to support large simulation experiments.

Exercises

  1. Install the Chi programming environment at your computer.

  2. Test your first program.

    1. Construct the following program in a project in your workspace:

      model M():
    writeln("It works!")
end

    2. Compile, and simulate the model as explained in the tool manual (in Compile and simulate).

    3. Try to explain the result.

  3. Test a program with model parameters.

    1. Construct the following program in the same manner:

      model M(string s):
    write("%s\n")
end

    2. Simulate the model, where you have to set the Model instance text to M("OOPS") in the dialog box of the simulator.

    3. Try to explain the result.