In recent years Earth‑observation satellites have become smaller and more numerous. The result? A massive stream of images and measurements that must be downlinked to Earth. Low‑Earth‑Orbit (LEO) constellations can act as a "highway" for this data, but managing it is trickier than it looks.
The issue, called DTIC (Data Transmission in Communication Constellations), splits into two steps: first the data hop between satellites (inter‑satellite routing), then they reach a ground station. Researchers realized this scenario mirrors a production‑scheduling problem known as Hybrid Flow Shop, with a twist: ground antennas can handle several packets at once, as long as two‑dimensional space limits are respected.
To tackle it they introduced a new model, HFSP‑2D, and two algorithms. The first, NEHedd‑2D, quickly builds a basic solution. The second, TSMA (Two‑Stage Memetic Algorithm), starts from that base and refines it using two "job‑swapping" operators: one between different antennas (IPJS) and one within the same antenna (IAJS). Tests on small and large instances show that, when seeded with NEHedd‑2D, TSMA finds the optimal solution for tiny cases and outperforms all benchmark methods for larger ones. In practice, these tools could speed up and stabilise the flow of data from our orbital “home” to the planet, enabling sharper weather forecasts, finer environmental monitoring, and fresh scientific breakthroughs.
