Problem 46618. Kaggle 2020 Drone Delivery Contest
The 2020 Kaggle Drone contest is an optimization task to maximize net customer satisfaction by using 30 drones across 10 warehouses to fulfill 1250 customer multi-item, 400 distinct items(products), orders. Satisfaction is (1-delivery_time/max_time)*100 and 0 if delivery not completed by max_time. The max time of 112993 is easily beaten with typical worse time of 40K.
This contest subset has disabled moving items from warehouse to warehouse thus wait times are not used.
The maximum score is 125000. To succeed as a DroneManager requires a score of 110K, 5th at Kaggle contest 9/26/20.
Input: [rows,cols,numdrones,maxturns,maxDronewt,numproducts,numOrders,delivery_xy_qty,delivery_list,distance_delivery&warehouse_to_delivery&warehouse, distance_warehouse_to_delivery,permutation_cell_array_for1to9]
Output: Commands matrix [number of commands,5] The number of commands is likely to be 18K to 20K.
Loading from a warehouse: [drone# 1 warehouse# item# quantity]. Drone1:30, Warehouse1:10, Item1:400. The 1 is LOAD.
Only one item type can be loaded on to drone at a time. Each Load/Deliver command consumes 1 unit of time.
Delivery for an order: [drone# 3 delivery# item# quantity]. Drone1:30, Delivery1:1250, Item1:400. The 3 is Deliver.
The final delivery time for an order is the latest drone time inclusive of final delivery time unit.
Additional approach comments are at Kaggle Drone 111401 and in the template along with using a provided routine to create a Kaggle python submission file.
Delivery/Warehouse Map. Warehouses red*, Single item delivery redO, Two item delivery blackO, >2 items greenO
Solution CommentsShow comments
Problem Recent Solvers1