Mixed integer programming & heuristic scheduling for space communication networks

We introduce mixed integer programming and heuristic scheduling for space communication networks in this paper. The considered communication network consists of space and ground assets with the link dynamics between any two assets vary with respect to time, distance, and telecom configurations. One asset could be communicating with another at very high data rates at one time and at other times, communication is impossible, as the asset could be inaccessible from the network due to planetary occultation. Based on the network's geometric dynamics and link capabilities, the start time, end time, and link configuration of each view-period are optimized so that the resulting schedule meets the objectives and satisfies the operation constraints imposed by the missions. Mathematical framework for the constrained optimization along with the software, Mars Relay Network Planning Tool (MRNPT), was developed in 2003 at Jet Propulsion Laboratory. In this paper, we broaden the study to include communication networks whose nodal elements can simultaneously support multiple spacecrafts and vice versa. Sample scenarios include multiple combinations of ground stations at a DSN site can form an antennas array to track a spacecraft and conversely a single ground station can support multiple spacecrafts simultaneously. Such unique network capabilities make the schedule optimization problem more challenging. Namely, the number of passes between a spacecraft and a set of ground stations soars from linear to exponential. As a result, the search space for the network optimization using start times and end times grows intractably. The new framework, casted in this paper as a mixed integer problem, provides the means to keep the search space dimensions to the linear order. The difficulty is now in the court of the optimization process. Traditional methods such as the Sequential Quadratic Optimization (SQP), used in MRNPT, are time-consuming and do not guarantee a global solution. In this paper, we propose to solve the constrained optimization problem in two phases. The first phase uses heuristic methods such as the ant colony method, particle swarming optimization, and genetic algorithm to seek a near optimal solution among a list of feasible initial populations. The final optimal solution can be found by using the solution of the first phase as the initial condition to the SQP algorithm. We demonstrate the above problem formulation and optimization schemes with a large-scale network that includes the DSN ground stations and a number of spacecraft of deep space missions.