Time-varying minimum-cost portfolio insurance problem via an adaptive fuzzy-power LVI-PDNN

Citation:

Katsikis, V. N., Mourtas, S. D., Stanimirović, P. S., Li, S., & Cao, X. (2023). Time-varying minimum-cost portfolio insurance problem via an adaptive fuzzy-power LVI-PDNN. Applied Mathematics and Computation, 441, 127700. Copy at http://www.tinyurl.com/29p9qe4f

Abstract:

It is well known that minimum-cost portfolio insurance (MPI) is an essential investment strategy. This article presents a time-varying version of the original static MPI problem, which is thus more realistic. Then, to solve it efficiently, we propose a powerful recurrent neural network called the linear-variational-inequality primal-dual neural network (LVI-PDNN). By doing so, we overcome the drawbacks of the static approach and propose an online solution. In order to improve the performance of the standard LVI-PDNN model, an adaptive fuzzy-power LVI-PDNN (F-LVI-PDNN) model is also introduced and studied. This model combines the fuzzy control technique with LVI-PDNN. Numerical experiments and computer simulations confirm the F-LVI-PDNN model’s superiority over the LVI-PDNN model and show that our approach is a splendid option to accustomed MATLAB procedures.

Website