Dutch Seminar on Optimization (online series) with Esteban Gabory (CWI) and Donato Maragno (UvA)

The Dutch Seminar on Optimization is an initiative to bring together researchers from the Netherlands and beyond, with topics that are centered around Optimization in a broad sense. We would like to invite all researchers, especially also PhD students, who are working on related topics to join the events. We hereby announce the following two talks, given by PhD-students:

Speaker: Esteban Gabory (CWI)
Title: On Strings Having the Same Length-k Substrings
Date: Thursday 31st March, 4pm CET

Abstract:
Let Substr_k(T) denote the set of length-k substrings of a given string T for a given integer k>0. We study the following basic string problem, called Shortest equivalent string: Given a set S_k of n length-k strings and an integer z>0, find a shortest string T such that Substr_k(T)=S_k. The Shortest equivalent string problem arises naturally as an encoding problem in many real-world applications; e.g., in data privacy, in data compression, and in bioinformatics. We answer this problem by showing that it is equivalent to finding a shortest walk that crosses every edge of a graph, and we give an algorithm for the latter problem.

https://arxiv.org/abs/2109.13566

Speaker: Donato Maragno (UvA)
Title: Mixed-Integer Optimization with Constraint Learning
Date: Thursday 31st March, 4:30pm CET

Abstract:
We establish a broad methodological foundation for mixed-integer optimization with learned constraints. We propose an end-to-end pipeline for data-driven decision making in which constraints and objectives are directly learned from data using machine learning, and the trained models are embedded in an optimization formulation. We exploit the mixed-integer optimization-representability of many machine learning methods, including linear models, decision trees, ensembles, and multi-layer perceptrons. The consideration of multiple methods allows us to capture various underlying relationships between decisions, contextual variables, and outcomes. We also characterize a decision trust region using the convex hull of the observations, to ensure credible recommendations and avoid extrapolation. We efficiently incorporate this representation using column generation and clustering. In combination with domain-driven constraints and objective terms, the embedded models and trust region define a mixed-integer optimization problem for prescription generation. We implement this framework as a Python package (OptiCL) for practitioners. We demonstrate the method in both chemotherapy optimization and World Food Programme planning. The case studies illustrate the benefit of the framework in generating high-quality prescriptions, the value added by the trust region, the incorporation of multiple machine learning methods, and the inclusion of multiple learned constraints.

Joint work with: Holly Wiberg, Dimitris Bertsimas, S. Ilker Birbil, Dick den Hertog, Adejuyigbe Fajemisin

Paper:
https://arxiv.org/abs/2111.04469
Code:
https://github.com/hwiberg/OptiCL