Operations is Child’s Play

What is your current capacity utilization?

How do you maximize profit when your excess capacity cannot be carried over to next period?

How can you predict future demand?

How do you know how much raw materials to order for each period?

If you know the past demand history, you can likely predict the future.

Kids these days learn this just by watching Cyber Chase on PBS, particularly one specific episode, aptly titled as “Past Perfect Prediction“. The kids run a some sort of oil change garage to raise money. They have to order Cryoxide, the raw material that costs $15 a can and expires at the end of the day. They charge $32.5 for the service. They place an order for Cryoxide the previous day and it gets delivered in the morning.


Past Perfect Prediction

Convinced that the last piece he needs to activate his powerful new machine is hidden in Slider’s garage, Hacker threatens to evict the teen unless he pays up on an old debt. Enter the kids and Digit. As a way to raise the money, they convince Slider to open the garage for business – just like his dad did. They do, but quickly discover that there’s more to it than meets the eye. Can they unlock the past to find the key to saving Slider’s future?

On the first day they order 66 cans based on one receipt they find in their father’s files. As it turns out they could use only 30 of the cans, wasting the other 36 cans. They figure out that that was just one data point and it was also from a Saturday whereas they started work on Monday. Their initial search gives them one past receipt for each day of the week. Not satisfied with the dataset they search more and find the receipts for the whole month. They find the average demand for each day and place a order for each remaining day of the week.

Perfect. They end up using every can everyday and end up making a wheelbarrow load of money.

Note: Surely you don’t think kids can handle standard deviation and normal distribution do you? It is okay if they left that out and simply described it as average over the month. But at least they showed them the distribution is bimodal (weekdays and weekends)