No matter how perfect your data, how detailed your model, it is hard to compete with the expertise of experienced planners and dispatchers.
While black box models have their place, our standard approach to planning, whether you are planning for next year, tomorrow or right now, is to keep the human in the loop.
For real-time models, we can provide a ranked list of recommended actions, starting with the decision we think is best, but allowing for the possibility that a knowledgeable dispatcher may know something that requires the use of alternative recommendations. Whether we are recommending the location for a spare part, the best person to do a job, or the best equipment, we believe in the power of providing choices. And if our best recommendation is not used, we update the plan immediately.
For strategic models, we allow the user to review and override decisions, at different levels of granularity. A planner may want to provide broad guidance (maintaining a minimum number of spares in a region, limiting the flows in a traffic lane). These overrides are recognized and incorporated seamlessly into the simulation, providing fast feedback for a smooth, interactive planning session.