A potential client came to us in their search for simulation software; they wanted to understand how SIMUL8 could be applied to critical path analysis within a production machine that had many complex parallel steps. If they could easily understand the critical path elements, they could focus their attention on the best process steps for overall cycle time reduction. In this example, all process times are deterministic (fixed distributions).

One of our clients has a production line in France that they intend to move to Romania. They have started to develop a simulation model to capture and document the inner workings of the line including cycle times, line efficiencies, labor assignments, station capacities and performance metrics.

They will then unistall, pack-up, and move the entire line 2500km East where it will be reassembled, configured, and production resumed.

The goal of the simulation is mainly to transfer knowledge from the old to the new plant to ensure that production targets are maintained. Normally, we see simulation used to identify and test changes to a system in order to improve it. This time, the simulation model will be used to ensure that everything remains the same.

This is a common situation. As a simple example, consider a receptionist whose primary duties are to answer phone calls and emails. Occasionally, the receptionist must also leave to fax a document or make copies. The best way to handle this depends upon how the employee should respond when an interruption occurs. Will the employee first complete the current task, then deal with the interruption or will the employee immediately drop the primary activity (even if in the middle of a task)?

The following SIMUL8 feature is very convenient method of randomly consuming a resource's time without needing to explicitly model the arrival and process times of the interruption.  If you are interested in the more detailed and complex method of modeling the interruption, please see Work Center Interruption.

The simplest approach is to use the availability feature of the resource in question.  Much the same as a work center's efficiency, availability randomly creates resource absences.  You can achieve this by either using the auto field by entering an availability % and  average absence time, or by using the detailed option.

The detailed option allows you to enter more accurate distribution parameters for time between absences and time to return.

The changeover requirement is detected as a work item enters a work center. At this moment, the work center can check label values, check the number of items processed (or the amount of working time) since the last changeover and determine if a changeover is required before starting the next work item.

Since changeovers were introduced, SIMUL8 added:

• the changeover result so we could evaluate what percentage of time a work center spent in changeover
• resource requirements for changeovers
• a Visual Logic event at the end of changeover (On State Change)

These additions have all been great, but we are still left with the fact that a changeover occurs at the beginning of a work center's process time. This means we need the next work item to trigger the changeover. But what if we want to use changeover to model cleanup or some other post-process time event?

Our frequent work in healthcare provides several good examples of this situation. Sometimes a pre-process changeover just can't accurately reflect the true process. Imagine that a trauma center's operating room has been left a mess from the previous patient. Would the staff wait until the next trauma patient arrives to clean and prep the room? Of course not. To do so would mean that the staff has both failed to make use of the free time between patients and, worse, risks making a potentially critical patient wait to receive care. Instead, staff would initiate the cleaning process as soon as the previous patient has left the room. 

So how can we model this in SIMUL8 given that changeover times are tied to the next patient by default? Simple - we trick the work center into initiating the changeover sequence ealier. One easy solution is to create a "change over" work item to pass through the work center after each "patient" is compete.  Using the On Work Complete Visual Logic, simply use the "Add work to Queue" visual logic command to create and add a new work item to the appropriate queue. Next set a label on the newly created work item to indicate that it is a "changeover" work item.

When a work item arrives into the work center use Visual Logic After Loading Work to determine if the item represents a "patient" or a "changeover".  If it is a "Patient", set the work center's operation time accordingly and the changeover time to zero. If the item is a changeover, set the operation time to zero and the changeover accordingly. When the item is complete, route the "changeover" to an exit point, and the "patient" to the next step in the process.

This solution will ensure that any changeover (i.e. cleaning) required will occur immediately after the previous process time completes. It also has the advantage of keeping the working percent and changeover percent results correct. You should keep in mind, however, that the number of work items completed result at the work center will reflect twice the number of patients, as it will have one count for each patient plus one count for each changeover.

This example illustrates a lesson that we have seen so often in our years of using SIMUL8 for complex problems - just because there may not be an immediate built-in dialog or check box to handle a given situation, that doesn't mean it can't be modeled with SIMUL8. With a little creativity and SIMUL8's powerful built-in programming capabilities, you can conquer just about any modeling challenge.  And don't forget that you can always ask us for help if you get stuck!

We have been working with and recommending our friends at Quetech since 2001 whenever we, or our clients, require assistance with time studies. Quetech is the developer of WorkStudy+. This software (for the Palm OS or Pocket PC) allows for streamlined and easy time studies using a handheld PDA.

Here's a quick run-down of how it works:

1. Develop your time study template by following an extremely user-friendly interface in Microsoft Excel.
2. Transfer the template to your handheld(s) using a USB cable.
3. Manually observe your process using a touch pen to record the start, end and category of tasks as well as many other customizable fields. Repeat until you've completed your data collection plan.
4. Transfer the data back to Excel for reporting and built-in statistical analysis.

We recently coached BJC HealthCare through a manual data collection effort with the goal of understanding the resource requirements to dispense medication within a hospital. The BJC study was covered tasks performed by the nursing staff within each unit, as well as order fulfillment tasks performed by pharmacists and technicians. WorkStudy+(combined with a few inexpensive PDAs from Walmart) proved invaluable in the manual data collection effort.

If you find yourself faced with a significant manual data collection project, consider this cost-effective, efficient solution. The days of stop watches, clipboards, and data entry will certainly not be missed.



We all know how elevators traditionally work - you push an up or down button, wait for the first available car going in your direction, pile in with everyone else and push the button for your desired floor. The elevator then stops in sequence at each floor chosen, meaning you may have several stops before you get to your floor.

The Sheraton's new elevator system approaches the problem differently. Rather than asking passengers to make an up or down selection as a first step, this system requires passengers to select a floor number before they even get onto a car. Once inside the elevator, passengers can no longer make a floor choice. Instead, they only see a list of potential stops.  Moving this critical 'which floor' information one step forward in the process allows the elevator system to group passengers by floor, thereby avoiding the 'milk run' to multiple floors required by traditional elevator set-ups. 

With the new system, as with the old, there is no guarantee that a car will be available for you the instant you want it. But with the intelligent grouping of passengers, you can expect a shorter ride to your final destination.

To a simulation guy who has made a career out of trying to improve systems, this seems revolutionary. I have seen banks of elevators grouped for low floors and high floors, or penthouse only elevators, but I have never seen a truly intelligent elevator system like this one. Once again, maybe I just don't get out enough.

As a simulation instructor, this stuck me as not only a smart system, but also a great learning exercise. So, we are offering a $50 gift certificate to the person who can create the best simulation(s) to compare a traditional elevator with this new "batch" elevator.  If you are interested, please send us an email using the contact form and we will forward you the parameters and assumptions for this little competition.

Here's how the feature works: You allow a work center to dynamically set its resource requirements by referencing a label on the work items bound for the work center. If the value of the label is 1, then the work center requires the first resource in the defined list of resources.  If the value of the label is 2 then the second resource would be required, and so on.

Because the Select Resource by Label feature does not support situations where multiple resources are required, such as a task that requires both a doctor and a nurse, I could not think of an immediate use for this new feature.

Last week, that changed. I had the pleasure of teaching SIMUL8 to a six sigma black belt from a major mortgage lending company. We found that the Select Resource by Label feature would nicely solve a staff allocation issue that would have previously required messy Visual Logic. 

In the situation that came up last week, a new mortgage application must be assigned to an individual mortgage 'specialist'. Once assigned, the 'file' belongs to the same specialist for the duration of the lenghtly application review process.  If we were to model the specialists as a single skill and set the number of resources available equal to the total number of specialists available, the total time in system would not be accurate as the 'pooling' would inaccurately portray capacity.  If Speciality A was not available, all of their files would sit and wait. Specialist B, C, D... would not be covering for them.

To solve this, we:

1. Created a specific resource for each specialist. 
2. Created one (replicated) work center for each manual task in the mortgage application process.
3. Assigned the 'file' to an individual Specialist via a label using a slighly complex 'circulate' assignment algorithm.
4. Used the Select Resource by Label feature to easily reflect which specialist was required by each file for each work center (task) in the process.

This approach offered a number of attractive advantages. It:

• Minimized the number of work centers in the simulation
• Minimized the amount of complex Visual Logic required
• Maintained a nice "Flow Chart" graphical layout of the simulation which enhanced the presentation value of the simulation
• Allowed for flexible and comprehensive staffing schedules
• Allowed for staff specific process times where warranted

Fitting distributions to data and why you are probably doing it wrong

by David Vose
Feb. 15, 2010

This discussion assumes that you already have a basic understanding of how to set and use labels. For more complete information on these topics, see our book Learning SIMUL8: The Complete Guide, or ask about our training programs.

Description:

An interruptible work center will accept higher priority work items in the middle of a cycle. Sometimes referred to as 'preempting,' this option causes the work center to set aside the work item it is currently processing to select the higher priority item, process it, then reselect the previous item where it left off. This act of preempting is a bit tricky to get to work properly.