Optimizing participant selection for fault-tolerant decision making in orbit using mixed integer linear programming

In challenging environments such as space, where decisions made by a network of satellites can be prone to inaccuracies or biases, leveraging smarter systems for onboard data processing, decision making is becoming increasingly common. To ensure fault tolerance within the network, consensus mechanisms play a crucial role. However, in a dynamically changing network topology, achieving consensus among all satellites can become excessively time consuming. To address this issue, the practical Byzantine fault-tolerance algorithm is employed, utilizing satellite trajectories as input to determine the time required for achieving consensus across a subnetwork of satellites. To optimize the selection of subsets for consensus, a mixed integer linear programming approach is developed. This method is then applied to analyze the characteristics of optimal subsets using satellites from the International Charter: Space and Major Disasters (ICSMD) over a predefined maximum time horizon. Results indicate that consensus within these satellites can be reached in less than 3.3 h in half of cases studied. Two satellites that are within the maximum communication range at all times are oversubscribed for taking part in the subnetwork. A further analysis has been completed to analyze which are the best set of orbital parameters for taking part in a consensus network as part of the ICSMD.