Matheuristics for offshore wind farms cable routing

A matheuristic approach is used to solve the inter-array cable optimization problem for offshore wind farms. The problem consists in finding an optimal cable route to connect all the turbines to one (or more) offshore substation(s). The model considers different constraints, such as cable capacity, prices, no crossings, a limited number of connections to each substation, and possible presence of obstacles in the site. First, a MILP model is defined. Even if it is well performing in general, in some particularly difficult instances it requires a large amount of time. Therefore we investigated three different matheuristics based on a relaxed version of the model. The heuristics developed are: - a random heuristic: we randomly fix some of the arcs of the current best solution and re-optimize the others using the MILP solver - a distance based heuristic: we randomly fix some of the arcs of the current best solution based on the distance to the substation(s) - a sector heuristic: we randomly decompose the problem into sectors, fixing the arcs outside the sector and re-optimize those inside Each of the three matheuristics has been tuned to find a proper trade-off between neighbourhood size and solution time. Finally, the different matheuristics are combined in order to define the best strategy to approach the overall optimization problem.