Overview / Usage

The problem has previously been approached by various optimization techniques like genetic algorithms, particle swarm optimization and various other heuristic methods. Since every satellite has panels of different sizes, different forbidden region sizes and different cell sizes, there is no absolute proof that these non-deterministic methods will provide optimal solution each time. This invokes a demand to develop a method that always promises an optimal solution. We have developed a convex optimization based algorithm that provides optimal solution. It is a two-stage algorithm; a heuristic based approach to convert our problem to a convex optimization problem and mixed integer convex programming to solve the convex problem for solar cells are of fixed size. The problem is optimized using CVX library, MOSEK and KNITRO solver.

​

Methodology / Approach

The problem has previously been approached by various optimization techniques like genetic algorithms, particle swarm optimization and various other heuristic methods. Since every satellite has panels of different sizes, different forbidden region sizes and different cell sizes, there is no absolute proof that these non-deterministic methods will provide optimal solution each time. This invokes a demand to develop a method that always promises an optimal solution. We have developed a convex optimization based algorithm that provides optimal solution. It is a two-stage algorithm; a heuristic based approach to convert our problem to a convex optimization problem and mixed integer convex programming to solve the convex problem for solar cells are of fixed size. The problem is optimized using CVX library, MOSEK and KNITRO solver

Technologies Used

Matlab and CVX

Other links

• DOI for my report for this project