Can You Solve Alibaba’s ‘Fastest Delivery’ Network‑Flow Challenge?

In mid‑September Alibaba launched a global mathematics competition to demonstrate the value of fundamental science, especially mathematics, and to let the public appreciate its beauty; AI is assisting the grading process.

The contest quickly attracted nearly 40,000 participants from top universities such as Harvard, MIT, Tsinghua and Zhejiang, setting a recent record for math‑focused competitions. Results will be announced in mid‑October with a final round in November.

Featured Problem: “How Can a Delivery Rider Deliver Fastest?”

The problem presents an undirected graph where numbers inside circles denote locations and numbers on edges denote distances.

Part a asks for the shortest route for a rider starting at point A who must pick up three orders, deliver them, and respect a vehicle capacity of two orders per trip.

Part b generalizes to a random graph where each traversed edge may, with a given probability, generate an additional order to the same destination; the question is the expected number of such extra orders and the probability of receiving at least one.

Part c turns the scenario into an optimization problem: each completed delivery yields a fixed revenue r, while the total path length is a cost (coefficient 1). With the same capacity constraint, the rider must choose a route that maximizes net profit, taking into account the chance of receiving extra orders without extra cost.

Why These Scenarios?

Organizers chose e‑commerce‑related settings—pricing and food delivery—because they are easy for the public to understand and directly illustrate how mathematical models can improve real‑world efficiency. The “fastest delivery” question is essentially a classic network‑flow problem; applying algorithms can increase rider efficiency by 15‑30 % and reduce carbon emissions.

Expert Insights

Mathematicians emphasized that such applied problems are rarely featured in traditional contests. By simplifying everyday logistics into mathematical models, participants can explore optimization, probability, and graph theory in a tangible context.

For those interested in solutions, Alibaba’s Machine Intelligence team provides official answer PDFs via a QR code and public account.