It’s easy enough to confuse automation with optimization. In both cases, you press a button, wait a few seconds, and whoops… there’s your result. But while the external actions look alike, the results are very different.
What automation really automates
Automation automates what human planners do. So what do planners do?
Planners oversimplify.
To cope with the overwhelming complexity of supply chain planning, planners base their decisions on simple rules of thumb or heuristics. For example, a planner who is assigning shipments to trucks is probably going to follow a rule such as ‘nearest next delivery’. When picking the next shipment to be assigned, the planner will look for the nearest delivery address to the previously assigned shipment.
It’s easy enough to automate such a rule, but the results will be disappointing for at least three reasons:
(1) The real world isn’t simple: The rule has ignored a whole host of constraints, rules and regulations, customers preferences – and the list goes on.
(2) It’s not the best plan because many options (in this case, better routes) have not been considered.
(3) You have no idea how the plan affects your KPIs.
The advantages – and limitations – of optimization
Optimization considers a huge array of possible combinations to arrive at a result that no human planner could ever hope to achieve. It incorporates all constraints, business rules, regulations and preferences; and enables planners to optimize plans based on the KPIs that need improving.
In an ideal world, optimization would be enough. In the real world of changing circumstances and constant disruptions, optimization needs to be supplemented with something else.
Suppose a driver is delayed for 30 minutes and is going to be late for the next delivery. What should be done? Should the driver skip the next delivery? Or should he or she make the next delivery in spite of the knock-on effect this will have on subsequent deliveries?
In situations where the disruption is relatively small, global or even local re-optimization would be an overkill. Any small improvement in KPIs would be outweighed by the pain of re-optimizing the original plan.
What planners need is an intelligent system that will help them grasp the impact of the disruption, and offer suggestions on how to solve any problems that arise.
This kind of decision support helps planners respond effectively to changing circumstances by:
– Calculating all the consequences of a disruption or change of plan
– Flagging problems that need to be resolved
– Suggesting solutions
– Highlighting how the decisions will affect KPIs
Optimization and decision support: the combination that’s ‘real-world ready’
It’s no surprise then that combining optimization with decision support produces the best results. You get the advantages of optimization, while ensuring that planners can step in and tweak plans when necessary.
After all, the best way to ensure that an optimized plan can be implemented is to make sure it can be revised.
