Paperlane

no. 12 · Tuesday 10 November 2026

Learning from the past in an irreversible investment problem

Topias Tolonen-Weckström

math.OC · v3

Abstract

We consider an irreversible investment problem under incomplete information, where the investor decides whether and when to make investments in a project. Upon investment, the investor acquires previously hidden information from the project's past (''learning from the past''), and so the learning rate of the problem is controlled by investing. We set up this original problem as a recursively defined stopping problem with a learning-from-the-past effect. We show that, under general conditions, at each recursion step there exists a one-sided stopping boundary. We proceed to present the optimal investment strategy as a sequence of semi-explicit stopping boundaries derived from the smooth fit conditions of the corresponding free-boundary problems. Feasibility of our approach is then demonstrated by solving example boundaries numerically and by performing sensitivity analysis with respect to different model parameters.

comments
To appear in: Journal of Applied Probability
fields
Optimization and Control
pdf ↗ page arXiv:2508.21731 ↗