Article: What if comes to life

SAMIR KHAN has been looking at Micro Saint, a powerful tool for simulating discrete events and assessing different outcomes.

You’re the production manager of a factory. Piles of intermediate materials building up at various points in your manufacturing process. You want to know whether the benefit of less work in progress with a Just-In-Time process is worth the increase in production time.

A hospital manager wants to find the optimum staffing levels and distribution of equipment in an emergency ward given constraints on the amount of money that’s available.

A helicopter designer wants to know if a pilot can process the shear amount of sensory data available and still be able to control the aircraft.
What software package did these people turn to? Micro Saint. It’s a tool for simulating discrete events and can be applied to all the above "what-if" situations and more.

The first step is to define your process by developing a flowchart. This is as simple as placing tasks on a network and then connecting the tasks with execution paths. You then define the logic behind your process: you can state that the milling of drill bits can only start if there’s a lathe and a technician free as well as enough tempered steel waiting to be processed.

A task can have more than one outcome. A hospital manager may know from experience that 95% of patients that walk through the door have light injuries, while the other 5% require a greater degree of care ‹ that’s easily modelled in Micro Saint. The 5% who need immediate attention can be routed to specialist treatment rooms while the luckier 95% have to queue for a doctor outside the minor treatment rooms. Those waiting for minor treatment can be assigned a number that indicates the degree of injury. Patients with a number higher than a set value can be rerouted for further treatment, while the rest can go home.

As well as the probabilistic and tactical decisions described above, a task can also have multiple outcomes. An order for a car can spawn two separate production lines - one that constructs the chassis and another that fabricates the interior. The production lines then combine to form the final car.

Life’s full of grey areas and that’s reflected by Micro Saint. You can specify probability distributions for the time it takes a person to walk from one end of a production line to another. The production manager mentioned above can use Optquest, a module that’s packaged with Micro Saint, to funnel the search for the optimum manning levels given constraints on money and limits on operator workload.

Running your simulation brings the network to life: patients move from reception to treatment rooms, queues of calls build dynamically, waiting to be answered by the next free receptionist. You also get the option of collecting data, such as plotting the number of free technicians as a function of time or extracting the maximum number of patients that queue for a free doctor during the day.

Actionview, another bundled module, lets you build an animation of your process, so you can graphically see whether a technician is available at a particular point in time. It’s a great tool for presenting your simulation to a wider audience. And if the power of Micro Saint isn’t enough, you can access the functionality of other software tools with COM services. With a programming language acting as the middleware, Micro Saint can talk to another application, even while the simulation is running.