Tomohiro Koana

Email: tomohiro.koana@gmail.com

About Me

I am a Project Assistant Professor (特任助教) at the University of Tokyo. My research is in theoretical computer science, especially the design and analysis of algorithms for NP-hard problems. I am currently interested in algebraic methods for faster exponential-time algorithms, parameterized and exact algorithms, graph and matroid algorithms, counting problems, derandomization, and time-space tradeoffs.

Education

Technische Universität Berlin

PhD in Computer Science

February 2020 - January 2024

Thesis: New Parameterized Algorithms for Matching and Packing Problems.
Advisor: Prof. Rolf Niedermeier.

Technische Universität Berlin

MSc in Computer Science

October 2017 - December 2019

Thesis: Algorithmic and Structural Aspects of Matrix Completion Problems.
Advisor: Prof. Rolf Niedermeier.

The University of Tokyo

BEng in Information and Communication Engineering

April 2013 - March 2017

Thesis: An analytic approach for statistical properties of stochastic oscillators and its application.
Advisor: Prof. Yoshihiko Hasegawa.

Selected Publications

See DBLP or Google Scholar for complete publication lists.

Determinantal Sieving

Eduard Eiben, Tomohiro Koana, Magnus Wahlström.
TheoretiCS 2025 | SODA 2024

Introduces determinantal sieving, an algebraic framework for parameterized algorithms over linear matroids. The method extends multilinear detection to matroidal constraints and gives cleaner or faster algorithms for a range of problems.

Faster Edge Coloring by Partition Sieving

Shyan Akmal, Tomohiro Koana.
STACS 2025

Gives an exact algorithm for Edge Coloring that improves over the standard 2^m-time baseline while using only polynomial space. The result is a step toward faster exact algorithms for one of the classical NP-hard coloring problems.

Induced Matching Below Guarantees: Average Paves the Way for Fixed-Parameter Tractability

Tomohiro Koana.
STACS 2023

Studies below-guarantee parameterizations of Induced Matching. The paper shows that two natural smaller guarantees are unlikely to yield fixed-parameter tractability on their own, but that their average does lead to an FPT branching algorithm using Gallai-Edmonds decomposition.

The Complexity of Finding Fair Many-To-One Matchings

Niclas Boehmer, Tomohiro Koana.
ACM Transactions on Algorithms 2024 | ICALP 2022

Analyzes fair variants of bipartite many-to-one matching where colors model attributes. The paper proves hardness even under strong restrictions, gives FPT algorithms for structural parameters, and uses integer programs based on Hall-type conditions and a separation theorem inspired by Frank’s work.

Exploiting c-Closure in Kernelization Algorithms for Graph Problems

Tomohiro Koana, Christian Komusiewicz, Frank Sommer.
SIAM Journal on Discrete Mathematics 2022 | ESA 2020

Applies the c-closed graph model to kernelization for classic graph problems. The key ingredient is a polynomial Ramsey-type bound for c-closed graphs, which yields fixed-parameter algorithms and polynomial kernels for problems such as Dominating Set and Induced Matching.