Solvers

Near production numerical solver flow.

THALES + PASQAL flow

From requests to Maximum Independent Set

From previous discretization in time of satellites data take opportunities, we end up with DTOS and PRIORITIES. We now need too feed this data to our solvers.

from spot.mis.solver import Solver
solver = Solver()

First, we need to collect opportunities from DTOs of each satelllite.

collected_opportunities = solver.collect_opportunities(dtos, priorities)

This array is then converted to a graph with edges constraining missions to respect operational constraints.

solving_graph = solving_map.graph_from_collected_opportunities(collected_opportunities)

These graphs can be fed to Maximum Independent Set solver

Networkx solver

We will need a classical solver to benchmark our computations on real hardware. We here describe the algorithm taken as state of the art resolution method to cover the MIS problemm.

classical_solver = NetworkxSolver()
mis_nodes = classical_solver.get_maximum_independet_set(solving_graph)

This returns and list of nodes which should be independent.

Rydberg solver

We can use already described PASQAL solution

from spot.mis.rydbgerg.qubo import SpotQuboSolver

quantum_solver = SpotQUBOSolver()
mis_nodes = quantum_solver.get_maximum_independet_set(solving_graph)

Or use a standalone autoencoder-driver solution (less reliable)

quantum_solver = PulseSolver()
mis_nodes  = quantum_solver.get_maximum_independet_set(solving_graph)

Recover global plan

Finally, independently of the quantum algorithm chosen, we can recover the global plan and compute metrics.

global_plan = solver.get_global_plan(mis_nodes)

Go to the next page (numerical experiments) to know more about how to interprete the results.