meta data for this page
Using JESS Client with DesignBuilder
Note: the JESS Client will be integrated into the DesignBuilder software from version 4. The information below applies only to certain beta versions of DesignBuilder V3 software in which the DB Job Server function is available.
Through the new JobServer function provided in the latest DesignBuilder Beta package, it is possible to send simulation jobs, especially those of the optimisation process, to the JESS server on the DMU cluster. This page describes how to setup the necessary components, and connect to the JESS server.
You will need
- DesignBuilder V3.1.0.089 or higher, you can download it from DesignBuilder.co.uk
- Java Runtime (JRE) v1.7.0 or higher, download it from oracle.com
- JESS Client for DB package, download it from Get jEPlus+
- A JESS account, you can ask for access by sending an email to firstname.lastname@example.org
- Good, unlimited internet connection, as running simulation remotely may generate high internet data traffic.
Some preparation steps have to be taken. These may have already been done, in which case you can move to the next section.
- Install DesignBuilder V3.1.0.089 (or higher)
- Check DesignBuilder's JobServer function is working correctly. Here are instructions on how to check the JobServer.
- Install Java Runtime. If you are not sure whether Java is already installed on your computer, follow these steps to check:
- Bring up the command prompt window. If you don't know how, look here
java -version, followed by the Enter key
- If you see something like this screenshot, everything is fine.
Initialize JESS account
Once you have downloaded the JESS Client for DB and received a JESS account, you must first perform the following steps to initialize the connection.
- Unpack JESS Client for DB in to a folder, for example, in
- Open the folder, you will see it contains one sub-folder named “EnergyPlus”
- Go inside the
EnergyPlusfolder, and then double-click on the Windows Batch File named
Register_Userif extension is hidden)
- The JESS Client GUI (see screenshot) will show. If the GUI does not appear, it is most likely that your Java Runtime has not been installed properly. Go back to the preparation steps.
- Please read carefully the legal documents (accessible by clicking on the “Open JESS Service Terms” button).
- If you have understood and agreed to the terms and conditions, please tick the corresponding checkbox
- If the connection is initialized successfully, a confirmation message will read “Session is valid” (see screenshot). You can now close the JESS Client GUI.
- If you cannot login, please contact us at email@example.com
Install JESS Client in DesignBuilder
- Now, open this folder in your file explorer if you are using Windows 7:
C:\Users\Yi\AppData\Local\DesignBuilder\JobServer\Jobs. On Windows XP, the folder may be
C:\Documents and Settings\Yi\Local Settings\Application Data\DesignBuilder\JobServer\Jobs. Remember to replace “Yi” with your own user name on the computer. If you cannot find such folder, it may be set to hidden, in which case you can copy and paste the path into the address bar of file explorer
- Rename the existing “EnergyPlus” folder to something else, e.g. “EnergyPlus.bak”
- Copy the “EnergyPlus” folder in the downloaded JESS Client for DB package into this location
- That's it! The screenshot shows how the folder looks after these steps.
Configure DB JobServer
With the current implementation of the DesignBuilder JobServer, there are three parameters that may affect how simulation jobs are executed in parallel. These are in the JobServer settings, Batch settings, and the population size of the Genetic Algorithm.
Manage Server window
The screenshot on the right shows the Manage Server window (see here for how to get to it). The “Concurrent jobs” field governs the maximum number of jobs the JobServer can submit to JESS concurrently. Please do NOT set this number higher than 20 in any case, as it may overload your computer and network.
Batch settings tab
This screenshot shows the “Number of jobs per batch” field in the Batch tab in Optimisation. Since simulation jobs of the optimisation process are submitted the DB JobServer in batches, this number governs how many jobs to be submitted the JobServer's queue in each batch. The optimisation process will not continue until all jobs in the current batch have returned. This number must be equal to or smaller than the population size explained in the next section.
Population size of GA
The optimisation algorithm (GA) used in DB progresses in consecutive “batches”, too, although a “batch” here is normally called a generation of solutions. The population size parameter governs the batch size of the GA process. Population size can be adjusted here on this window.
Important notes about remote simulation
Remote simulation means you are sending your model to the DMU servers via internet. This raises several issues and risks, such as security and cost. So, please read this section carefully.
Firstly, the remote simulation service we are offering is a limited trial, as part of the ADOPT research project. The trial period is from now till the closing date of the DB Competition (28 December). We may offer to extend the trial period till 31 March, 2013. Both the number of jobs you can submit per week, and the simulation time of each job, are capped. Currently, the maximum simulation time of each job is 60 minutes. Models take longer than 60 minutes to execute will fail. The maximum number of jobs you can submit each week will be determined later, depending on the workload of our servers.
This trial is “free”, which means we are not charging you for using the remote simulation service. However, using this service may not be completely cost free to you. Please consider your own cost for internet access, for example. DesignBuilder models may generate large amount of output data, which has to be transferred through the internet. If your ISP charge internet access by the amount of data you use, you must think very carefully before using this service. In any case, we strongly recommend that you select only the monthly and annual output level for optimisation purpose (see screenshot). This will help reduce data traffic.