Prescriptive analytics applications embrace most aspects of business operations. In fact whenever there is a resource allocation problem with constraining rules, many variables and well defined objectives, then it is likely that prescriptive analytics can be applied.
It’s application in marketing is gaining momentum, particularly combined with predictive analytics where customer response to various initiatives can be predicted. Resource allocation can then be optimized based on these predictions and the relevant business rules which determine how customers are to be treated. Of course optimization has been applied for decades to workforce and other resource deployment problems, but it is only now with much faster execution that other problems are becoming feasible. Even sports bodies are using optimization to maximize television ratings while catering to the needs of players and fans.
Optimization is used widely in the airline industry. Since margins are so thin frequent optimization can help turn unprofitable scenarios into profitable ones. Bad weather will mean re-optimizing resource allocation, and as seats are sold on a flight it is desirable to re-optimize the allocation of various classes of seat. Unscheduled maintenance can also cause resources to be re-allocated and optimization is essential if inefficiencies are to be ironed out of operations. In this particular application it is very important that optimization can be performed in less than 24 hours, and often in much shorter times.
The energy industries have used optimization for decades, but increased optimization speed and capability have meant new applications. With new energy sources, and particularly renewable ones, it is desirable to combine prescriptive analytics with predictive analytics to create long term energy scenarios. The predictive analytics are used to create long term energy forecasts and optimization is then applied to explore optimization of energy production.
An application ideally suited to optimization is that presented by vehicle rental firms. With thousands of vehicles assigned to hundreds of locations there is an obvious need to make sure that each vehicle is located where it will generate most revenue.
The earliest application of optimization included problems such as ‘least cost mix’. Here the aim is to produce a mix of component materials with a particular specification, given that each component has its own constituent properties. Animal food mix was the classic application where overall nutrient levels had to be met from mixing various ingredients, which each had their own nutrient levels. This and other similar problems are still solved using optimization technology, but look fairly primitive compared with the real-life problems that are being solved today.
Optimization technology will grow in use as it becomes more user friendly, executes with greater speed, and is more tightly integrated into the over systems environment. For many businesses optimization will provide an edge that cannot be achieved any other way.
Next article in this series: Prescriptive Analytics Strategy