CanaDAM 2019
SFU Harbour Centre, May 29 - 31, 2019

Computational methods in industrial mathematics I
Org: Bogumił Kamiński (Warsaw School of Economics, Poland) and Andrei Raigorodskii (Moscow Institute of Physics and Technology, Russia)

RICHARD DARLING, U. S. Department of Defense
Resolving labelling conflicts on big graphs via combinatorial data fusion  [PDF]

Transactional data streams attach labels to vertices in large graphs. When data comes from disparate and possibly inconsistent sources, such labels may be wrong, and the conflict is detectable in terms of a hypergraph of "forbidden sets". Combinatorial data fusion seeks graph cut algorithms to correct vertex labels with least cost. When graphs are very large, we implement quasilinear time approximations.

ŁUKASZ KRAIŃSKI, SGH Warsaw School of Economics
Optimization of road side units location within vehicle communication networks using multi-agent routing simulation  [PDF]

{\footnotesize co-authors: Przemyslaw Szufel, Bogumil Kaminski, Atefeh Mashatan, Pawel Pralat}

We consider an intelligent transportation systems (ITS) with vehicle-to-vehicle/infrastructure (V2V/V2I) networks. Using multi-agent approach, we simulate routing mechanism within V2I system governing city traffic and distributing information to agents. Communication infrastructure consists of Road Side Units (RSU) - transmitters limited in terms of data throughput and effective range which may lead to overload if sufficient number of agents is present in node. Under given constraints, RSU locations are optimized to provide stable communication in two scenarios: centralized (only V2I system) and hybrid (V2V clusters connected with V2I network).

MARCIN OPALSKI, SGH Warsaw School of Economics
Optimization of new roads construction by for intelligent transportation systems - an agent based spatial simulation approach  [PDF]

{\footnotesize co-authors: Przemyslaw Szufel, Bogumil Kaminski, Atefeh Mashatan, Pawel Pralat}

The goal of this work is to optimize routing within a vehicle-to-infrastructure (V2I) wireless communication using multi-agent approach. Actors (cars, traffic lights, road signs) communicate within each other and maximize collectively the through output of a heterogeneous, distributed Intelligent Transportation Systems (ITS). We simulate road conditions, determine the most congested areas and decide on where construction of new roads will reduce traffic jams. We will explain how the proposed approach can be further applied on a bigger scale.

PAWEL PRALAT, Ryerson University
Clustering via Hypergraph Modularity  [PDF]

Modularity is designed to measure the strength of division of a network into communities. Networks with high modularity have dense connections between the vertices within clusters but sparse connections between vertices of different clusters. As a result, modularity is often used in optimization methods for detecting community structure in networks. In fact, many important problems (including clustering) can be described using more general combinatorial objects, hypergraphs. Unfortunately, theoretical foundations as well as practical algorithms using hypergraphs are not well developed yet. Hence, we propose a hypergraph modularity function that generalizes its well established and widely used graph counterpart measure.