The optimization problem we study here consists in finding an optimal cable routing to connect a given number of offshore turbines to one (or more) offshore substation(s). Different constraints have to be respected, such as cable capacity, cable prices, crossing restrictions, limits on connections to substation(s), and possible presence of obstacles in the site. To solve this large-scale optimization problem we use a matheuristic approach, that is an hybridization of mathematical programming techniques and heuristics. First, a Mixed-Integer Linear Programming (MILP) model is defined. The MILP model is able to solve smaller instances to optimality but for large wind parks it fails in even finding a feasible solution. Therefore we investigate various matheuristics to handle this situation: the heuristics are used to decrease the number of variables in the optimization model by fixing some of them at each iteration. We propose and compare three different fixing strategy: “random fixing”, “distance based fixing” and “sector fixing”. Each of the three matheuristics has been tuned to find a proper trade-off between neighborhood size and solution time. Finally, we compare the solutions from the matheuristic framework with solutions from the initial MILP model on a number of real world instances, demonstrating the effectiveness of our approach when optimizing inter-array cable routing of big parks.
|Title of host publication||Proceedings of the 2016 European Control Conference (ECC)|
|Publication status||Published - 2016|
|Event||15th European Control Conference (ECC16) - Aalborg, Denmark|
Duration: 29 Jun 2016 → 1 Jul 2016
Conference number: 16
|Conference||15th European Control Conference (ECC16)|
|Period||29/06/2016 → 01/07/2016|