A uniform supplier has 24 items, next month's demand forecast, a purchasing budget and limited warehouse space. Which items should it reorder, and how many units of each? Move the constraints and the optimiser re-plans the order.
Where the dark line flattens, more money stops paying off, usually because warehouse space runs out first.
Spend by product type, optimised vs rule of thumb.
Expected lost and leftover units are averages across 80 demand scenarios for next month.
Forecasting says what demand might be. Prescriptive analytics decides what to do about it: here, order quantities that maximise expected profit.
Every unit ordered might sell, or might sit on the shelf tying up cash. Every unit not ordered might cost a sale and a customer. The model weighs both, item by item.
Budget, warehouse space, supplier minimum order quantities, and a fixed AED 60 cost for each item ordered.
A stochastic mixed-integer program in Python (PuLP + HiGHS), solving 80 demand scenarios for all 24 items at once in under a second.