Acceleration techniques for adaptive robust optimization transmission network expansion planning problems

The computational burden of two-stage adaptive robust optimization transmission network expansion planning problems increases when accurately representing the operation of power systems, i.e., when operational variability, inter-temporal operational constraints, and non-convex operational constraints are considered in the decision-making model. This motivates the use of acceleration techniques in order to avoid computationally intractable problems. The adaptive robust optimization problem is specifically solved using the nested column-and-constraint generation algorithm. We propose the application of two new acceleration techniques to this algorithm. On the one hand, the variables that model the uncertainty realizations are initialized to the solution obtained by using the alternating direction algorithm. On the other, the master problems of the solution procedure are relaxed by considering only certain cutting planes and including more constraints if the evolution of the bounds of the algorithm is not appropriate. Numerical results show that the use of the proposed acceleration techniques leads to reductions in the computational time of over 93%.