This work develops a mathematical formulation to determine the combinatorial search space of the Offshore Wind Transmission Optimization Problem (OWTOP). The model accounts for Capital Expenditures (CAPEX), Corrective Maintenance (CM), losses and Expected Energy Not Transmitted (EENT) and determines the optimal radial transmission system topology to connect a number of Offshore Wind Power Plants (OWPPs) to the shore. The model also considers the stochastic nature of wind. In this context a greedy search algorithm is developed capable of finding the globally optimal solution for the location of offshore substations (OSSs), sizing of the transmission infrastructure and topological layout. Rather than a single optimal solution, the algorithm finds a solution space of feasible topologies, bounded from below by the optimal radial solution. The algorithm is applied to 4 test cases including one based on the Belgian North Sea area. It is shown to outperform a traditional Transmission Network Expansion Problem (TNEP) formulation both in computational speed and solution quality.