Optimisation problems are by their nature extremely complex. For a start, external decisions impact optimisation problems. These external decisions are themselves influenced by other external factors, and so on, creating a rapidly ballooning level of complexity. Try to optimise transport, and you realise that the transport system is affected by the decisions of suppliers and customers, each of which is affected by other external factors. In theory, the system can be arbitrarily large, making optimisation almost impossible.
But that doesn’t mean we should give up. The issue is to determine which of those external factors are either arbitrary and/or have little to no impact on the problem in question. For example, consider delivery time windows. These can have a major cost impact on a transport provider, but often they are just there so that the customer knows when something will arrive; they could be set at any time. In the end, we can often put delivery time windows to the side.
A practical way to achieve this is to calculate the cost of any external constraints that could be changed. There's no point in looking at all constraints, since regulatory rules like statutory driver breaks and speed limits are fixed. But many external factors are variable, including delivery windows, delivery frequency and shift start times. All of these can impact cost and are worthy of some investigation. Finding out which of these influences are arbitrary and which can be leveraged for efficiency is a core part of the optimisation process. We need to undergo this process of sorting the grain from the chaff if we are to become more competitive.
