Design and plan for problems

 

It is traditionally known as Murphy's Law; if something could go wrong, it will go wrong.  

The inference from knowledge of this is to imagine what could go wrong, then plan to deal with that.  And here's where O.R. has a role to play.  Help people to make decisions, help them to evaluate strategies, help them to design better systems, and then help them to build protection around these pieces of advice.  This could be ways to prevent problems, or design ways to deal with them if they happen.

I was reminded of this essential lesson in a few ways over the last few days.  Tim Harford, writing for the Financial Times and the "i" newspaper, had a column "The power of negative thinking ", subtitled "Government could learn from city design about planning for the future ". He opened with the design of road signs which are increasingly being mounted on slip bases so that the uprights break away easily if hit by a vehicle.  If something could go wrong... then you don't want the accident to be made worse by entanglement with street furniture.

The subtitle refers to the U.K.  government's serious difficulties with test and trace for Covid-19, which failed many people because it used a database in Excel.  Never, never, never use Excel for a database!  And this was a critical set of data.  

As the article reports:  nobody with relevant expertise had been invited to consider the failure modes of the system.  No one asked if Murphy had tested the system.

Earlier this year, Tina and I bought a cycle carrier for the car.  Because Tina rides an electric bike, we needed a carrier that was able to support the extra weight of her cycle, so we had a towbar fitted and the carrier fits onto it.  We chose a towbar that could be removed, thinking that we would remove it in the winter when we wouldn't be taking the bikes out for the day.  What is wrong with this idea?  

Yesterday, we discovered.  Tina's bike had a puncture, six miles away from home, and on a cycle route, away from a public road.  That's a long way to push the bike.  On the spot repair is quite difficult, and it is advised that repairs are done by a cycle mechanic.  So, yours truly rode those six miles, lifted the carrier onto the towbar and came to meet Tina who had walked nearly three miles.  The weather wasn't very good, and I needed to change clothes when I got home because of the rain, and my top layer of waterproofs was in Tina's pannier.  

What can go wrong?  Punctures.  Bad weather.  Remoteness.  

Our response.  The towbar stays in place all the time.  If Tina has a puncture, yesterday's scheme works.  If I have a puncture, my bike will fit in the rear of the car so Tina can rescue me.  If the weather is dubious, we each have a pannier with our own waterproofs.   If need be, take a map.

And last week, another "What can go wrong?" experience.  Our wisteria hangs from a rope swag.  The climber traps water, which rots the rope.  After twenty years, the second rope rotted in rather less time than the first.  Now the wisteria is established, it is more inclined to trap water in the rope, and is heavier.  A third rope is needed.  We know what can go wrong.  So we react accordingly.  The new rope has synthetic fibers, which won't rot so quickly.  And, like most garden structures, we know that the wooden uprights will fail eventually.  But we also know that they won't be hit by a vehicle!