20.04.14

[Delivery Optimization] Clustering on area and delivery time

Doing delivery optimization with https://207inc.jp .This time, we will actually place the delivery destination on the map and simulate the actual data. The actual deliveries are many, but there are many duplicate addresses, so 136 addresses were prepared to take that into account. First, plot the address on a map.

This time, we used the clustering algorithm and performed the calculations with D-Wave Leap2. The clustering was done correctly by using a hybrid system that actually utilizes only the number of qubits for the number of clusters * number of delivery addresses. In this case, the clustering is sometimes wrong in the local simulator, so Leap2 is more accurate.

First, let’s take a closer look at the area for each cluster. And now we’re going to look at the time designation at the same time. At the time I assumed four types.

There are further time-specified delivery destinations within each cluster. Extract addresses for each individual time period.

This is how easy it was to do.



To COVID-19 Page