Mesquite Software

Home
Products and FAQ
Customers
To Order
Contact Us
Site Map
Documentation
News & Events

Purchase & Download
Or, log in here to
access your account.
 
Products and FAQ

CSIM for Java provides a comprehensive set of Java classes for implementing process-oriented, discrete-event simulation models. The software includes an extensive list of useful features that help programmers quickly build and efficiently run simulation models, described below.

  1. CSIM for Java's built-in simulation classes include classes for the following simulation structures:
    1. processes (CSIM processes) - a process is an active entity in a model
    2. facilities - a facility consists of one or more servers and a queue for waiting processes
    3. storages - a storage of a pool of tokens and a queue for waiting processes
    4. buffers - a buffer holds tokens; processes can get and put tokens into a buffer
    5. events - an event is used to synchronize processes
    6. mailboxes - a mailbox is used to exchange messages between processes
  2. The simulation classes listed above can automatically collect statistics on the usage of the object by processes.
  3. The simulation classes can be used in many ways, such as:
    1. In a model of a computer system, a facility can be used to model a CPU resource, possibly with multiple processors.
    2. In a model of a waiting room, a storage can be used to mimic the limitation on the number of waiting customers.
    3. In a model of a client-server system, a mailbox at the server can be used to collect requests arriving from the clients.
  4. CSIM for Java has a number of built-in functions for producing streams of (pseudo) random numbers. The probability distributions covered by these functions include:
    1. exponential(1.0) - negative exponential distribution with mean 1.0
    2. normal(5.0, 2.5) - normal distribution with mean 5.0 and standard deviation 2.5
    3. uniform(1.0, 10.0) - uniform distribution over the interval (1.0, 10.0)
    4. uniform_int(20, 80) - integer valued uniform distribution over the inveravl [20, 80]
  5. CSIM for Java has four objects that are used to collect statistics generated during the operation of the model:
    1. table - accumulate real values for summarization
    2. qtable - accumulate integer values; these values can reflect changes in the state of a resource, to, for example, summarize the queue length at a facility
    3. meter - accumulate data on the passage of processes past a fixed point in the system
    4. box - accumulate data describing both the number of processes in a fixed portion of the model and the time spent in that portion of the model.
  6. CSIM for Java contains useful features to facilitate obtaining statistically accurate results from a model:
    1. confidence intervals for tables and qtables; heuristics help eliminate bias caused by auto-correlated samples (e.g., response times)
    2. automatic run-length control - using a confidence interval, this feature lets a model execute until a pre-specified relative error for a confidence level is achieved, for the values collected in a table
    3. model reset - allows startup transients to be removed before data collection starts
  7. CSIM for Java provides a comprehensive set of reports on the usage of resources, delays by processes, and collected data. These reports come from:
    1. facilities, storages, and buffers
    2. monitored events and mailboxes
    3. tables, qtables, meters and boxes
  8. CSIM for Java provides "inspector functions" for each simulation object. These inspector functions can be used to track the state of the objects and to create customized reports on the usage of objects, delays by resources, and collected data.
  9. Models written in CSIM for Java are efficient. A thread pool minimizes overhead associated with creating and terminating simulation processes. In one test, using the J2EE Compiler in NetBeans, a model of an M/M/1 queue with server utilization of 0.5 required only 7.45 seconds (of CPU time) to simulate 50,168 completions. This test was run on a Dell Inspiron 2650 with a 1.7 GHz Pentium 4-M processor.
  10. The CSIM family of toolkits has been helping programmers build discrete-event simulation models for twenty years. The software is extremely stable and full-featured.
  11. CSIM for Java is attractively priced, with packages for professional developers, educators and corporate users.
  12. Comprehensive documentation is provided and includes a Getting Started documents and a User's Guide. You can access all CSIM documentation on our website.
More CSIM for Java Info

 

 
Home | Products/FAQ | Customers | To Order | Contact Us | Site Map | Documentation
© copyright 2005, Mesquite Software, all rights reserved.