In my day job, I’m paid to solve problems just as most people are, however what makes my job a bit unique is these problems happen to be very complex with tight time constraints. As these situations or incidents typically mean a company is either losing significant amounts of business, or there are health & safety issues to people and so on. This brings with it a fair amount of pressure to perform and have been the subject of previous posts like the discussion around decision trees and future ones regarding polymorphic methods.
However a situation of late got me thinking about a piece in the overall solution set if you will being the approach of selecting a serial solution set or parallel. As during a recent situation, had to make this call and while nothing new about being placed in that position, the call this time was “serial” which is unusual. While not something which I keep track of officially, would say that 9 out of 10 times the decision is “parallel” so why row upstream in this case?
This time was a little different in that there was line of sight to the solution and therefore the failure options were defined in a much smaller set. Failure set you ask, yes one of the first things which to do when faced with a “failure” if you will, is to develop a set of failure points be it items or attributes and then tie them together. This will then form your applied “graph theory”, while most of the time I can do these in my head the more complex ones get scribbled on paper and the “deeper” they run along with the high time pressure forces a “parallel” path to solution. However this time the depth was low (vertical representation of the graph map), almost flat while the width ran wide (horizontal representation).
With this there are couple points worth mentioning the first is applied “graph theory” and second is “Amdahl’s Law” as the understanding of both will help us in making decisions like these as each plays a complex role in helping us to find the solution we seek. Now before you go off and think I’m getting all academic on you here, I’m really not. The goal is to help understand how these tools can fit into your tool box and be built into your existing solution set. As you won’t have to whip out a slide rule or graphing calculator to use these as what’s important is to first understand these rules exist and second how they will factor in your decision process and add to your “decision tree” as discussed in a prior post.
To accomplish this task, we will break up the conversation into two additional posts where we will cover “graph theory” as what is it, to how can we apply it in order to optimize our path to a solution. Then we will look how our old friend Amdahl factors into the equation also as every time we make the call to take parallel path to a solution there is a “cost” intrinsic to doing so which we just can’t get away from…