Paragon Simulation Technical Article

“Fools rush in”

Vinod Bhatia

In my first job, if I wanted to write a report it was a really formal process – I’d write it out longhand, use the internal post to send it to the typing pool, and a few days later I’d get it back. This version would include misspellings, typing errors and maybe a space for me to manually add a formula, so the process could continue for two or three weeks. I’d then pass it to my section manager who would “red-line” my efforts i.e. correct my grammar, expose incomprehensible sentences, and highlight the sections of plain rubbish I’d written – with a thick red pen. No wonder every report seemed to take several months.

Today, technology allows me to make all the same mistakes, but in a fraction of the time. Worse, it allows me to release my unfinished and unchecked thoughts to a wide audience…

Trump targets ‘negative press covfefe’ in garbled midnight tweet, Washington Post, 31 May 2017

The technical aspects of modelling happily serve my nerdy nature, so it’s easy to become impulsive and reach for the keyboard and launch Excel as soon as the first hint of the simulation requirements is revealed. My intentions are all good (“I’m just playing around with simulation modelling concepts”), but before long I’m constructing complex expressions and then get distracted by some cool new feature that I haven’t spotted before (what’s a Sunburst chart?)

False start at a race

Experience shows that this is simply a false start. You may feel like you’ve been busy and accomplished something, but the inconvenient truth is that you’ll have to start over, once all the requirements are confirmed.

I would suggest that the first tool for “playing around with modelling concepts” is a pencil and notepad. Even the simplest Excel workbook will benefit from time spent in design, and it’s an absolutely essential step if the model needs to be built using a simulation tool, like Witness by Lanner® Group. The more complex the model, the greater the need to have an extended design phase that results in a document that is both complete and accurate.

In our simulation work the design phase is split into two: the Concept Design and the Technical Design. The Concept Design is usually a series of diagrams that document the process and how it is to be modelled, annotated with all the assumptions, simplifications and data requirements. There are two key aspects of the Concept Design; firstly it is not a Process Map (although it sometimes looks very similar); and secondly, it makes no assumptions about which technology will be used. This is so we can be free to focus on the business requirements of the model, and not on how it will be built. A separate decision is then taken on the technology to use, and when that is made the Concept Design is used to create the Technical Design.

The benefits of this approach are huge, particularly because there are always a few underlying assumptions in the design of the model. A poor Technical Design decision at the heart of a simulation model resulting from an incomplete or inaccurate Concept Design tends to be discovered late in model development, and will be time-consuming, frustrating and expensive to fix. But a good design phase results in a build phase that is significantly shorter, and produces a project that flows from start to finish without re-visiting earlier decisions. Perhaps the best outcome is that you are assured of a simulation model that is both complete and accurate – just like the design!

No false starts. No do-overs. And no “negative press covfefe” from your clients.