-
Two-Phase Optimization for PINN Training
Abstract: This work presents an algorithm for training Neural Networks where the loss function can be decomposed into two non-negative terms to be minimized. The proposed method is an adaptation of Inexact Restoration algorithms, constituting a two-phase method that imposes descent conditions. Some performance tests are carried out in PINN training.
Submitted 11 September, 2024; originally announced September 2024.
Comments: 8 pages
-
Optimal dividends and capital injection: A general Lévy model with extensions to regime-switching models
Abstract: This paper studies a general Lévy process model of the bail-out optimal dividend problem with an exponential time horizon, and further extends it to the regime-switching model. We first show the optimality of a double barrier strategy in the single-regime setting with a concave terminal payoff function. This is then applied to show the optimality of a Markov-modulated double barrier strategy in th… ▽ More
Submitted 24 October, 2024; v1 submitted 21 June, 2023; originally announced June 2023.
-
Effects of positive jumps of assets on endogenous bankruptcy and optimal capital structure: Continuous- and periodic-observation models
Abstract: In this paper, we study the optimal capital structure model with endogenous bankruptcy when the firm's asset value follows an exponential Lévy process with positive jumps. In the Leland-Toft framework \cite{LelandToft96}, we obtain the optimal bankruptcy barrier in the classical continuous-observation model and the periodic-observation model, recently studied by Palmowski et al.\ \cite{palmowski20… ▽ More
Submitted 24 August, 2020; originally announced August 2020.
-
arXiv:2006.03988 [pdf, ps, other]
Logarithmic correction to resistance
Abstract: We study the trace of the incipient infinite oriented branching random walk in $\mathbb{Z}^d \times \mathbb{Z}_+$ when the dimension is $d = 6$. Under suitable moment assumptions, we show that the electrical resistance between the root and level $n$ is $O(n \log^{-ξ}n )$ for a $ξ> 0$ that does not depend on details of the model.
Submitted 7 August, 2021; v1 submitted 6 June, 2020; originally announced June 2020.
Comments: 45 pages, 6 figures - Minor changes made to take into account the referee's comments
MSC Class: 60K50 (Primary) 60K35; 82C41; 31C20; 60J80 (Secondary)