Searcharxiv⌕ Search

arXiv subjects

Ivo Hedtke

Publications and source records attributed to Ivo Hedtke.

12 recordsLinked to original sources
arXiv2024

The SCIP Optimization Suite 9.0

The SCIP Optimization Suite provides a collection of software packages for mathematical optimization, centered around the constraint integer programming (CIP) framework SCIP. This report discusses the enhancements and extensions included in the SCIP Optimization Suite 9.0. The updates in SCIP 9.0 include improved symmetry handling, additions and improvements of nonlinear handlers and primal heuristics, a new cut generator and two new cut selection schemes, a new branching rule, a new LP interface, and several bug fixes. The SCIP Optimization Suite 9.0 also features new Rust and C++ interfaces for SCIP, new Python interface for SoPlex, along with enhancements to existing interfaces. The SCIP Optimization Suite 9.0 also includes new and improved features in the LP solver SoPlex, the presolving library PaPILO, the parallel framework UG, the decomposition framework GCG, and the SCIP extension SCIP-SDP. These additions and enhancements have resulted in an overall performance improvement of SCIP in terms of solving time, number of nodes in the branch-and-bound tree, as well as the reliability of the solver.

math.OC↗
arXiv2022

Partitioned vs. Integrated Planning of Hinterland Networks for LCL Transportation

Utilizing existing transportation networks better and designing (parts of) networks involves routing decisions to minimize transportation costs and maximize consolidation effects. We study the concrete example of hinterland networks for the truck-transportation of less-than-container-load (LCL) ocean freight shipments: A set of LCL shipments is given. They have to be routed through the hinterland network to be transported to an origin port and finally to the destination port via ship. On their way, they can be consolidated in hubs to full-container-load (FCL) shipments. The overall transportation cost depends on the selection of the origin port and the routing and consolidation in the hinterland network. A problem of this type appears for the global logistics provider DB Schenker. We translate the business problem into a hub location problem, describe it mathematically, and discuss solution strategies. As a result, an integrated modeling approach has several advantages over solving a simplified version of the problem, although it requires more computational effort.

math.OC↗
arXiv2021

Spot Market versus Full Charter Fleet: Decisions Support for Full Truck Load Tenders

This paper presents an approach to help business decision-makers gain market share by providing competitive tender offers for Full-Truck-Load (FTL) services. In particular, we compare operating a fleet of Full-Charter-Trucks (FCT), using spot-market (SM) capacity and a mixture of both options against each other. A Pickup and Delivery Problem is modeled, and solved using an Adaptive Large Neighborhood Search heuristic. Computational results indicate strong service benefits combining FCT and SM usage. Numerical experiments are presented in detail to support the findings. Additionally, a real-life case study originating from DB Schenker is presented.

math.OC↗
arXiv2018

Exact Algorithms for the Maximum Planar Subgraph Problem: New Models and Experiments

Given a graph $G$, the NP-hard Maximum Planar Subgraph problem asks for a planar subgraph of $G$ with the maximum number of edges. The only known non-trivial exact algorithm utilizes Kuratowski's famous planarity criterion and can be formulated as an integer linear program (ILP) or a pseudo-boolean satisfiability problem (PBS). We examine three alternative characterizations of planarity regarding their applicability to model maximum planar subgraphs. For each, we consider both ILP and PBS variants, investigate diverse formulation aspects, and evaluate their practical performance.

cs.DS↗
arXiv2013

A Fast Search Algorithm for Triple Product Property Triples and an Application for 5x5 Matrix Multiplication

We present a new fast search algorithm for Triple Product Property (TPP) triples as defined by Cohn and Umans in 2003. The new algorithm achieves a speed-up factor of 40 up to 194 in comparison to the best known search algorithm. With a parallelized version of the new algorithm we are able to search for TPP triples in groups up to order 55. As an application we identify a list of groups that would realize 5x5 matrix multiplication with under 100 resp. 125 scalar multiplications (the best known upper bound by Makarov 1987 resp. the trivial upper bound) if they contain a <5,5,5> TPP triple. With our new algorithm we show that no group can realize 5x5 matrix multiplication better than Makarov's algorithm.

math.GR↗
arXiv2011

Upgrading Subgroup Triple Product Property Triples

In 2003 COHN and UMANS introduced a group-theoretic approach to fast matrix multiplication. This involves finding large subsets of a group $G$ satisfying the Triple Product Property (TPP) as a means to bound the exponent $ω$ of matrix multiplication. Recently, Hedtke and Murthy discussed several methods to find TPP triples. Because the search space for subset triples is too large, it is only possible to focus on subgroup triples. We present methods to upgrade a given TPP triple to a bigger TPP triple. If no upgrade is possible we use reduction methods (based on random experiments and heuristics) to create a smaller TPP triple that can be used as input for the upgrade methods. If we apply the upgrade process for subset triples after one step with the upgrade method for subgroup triples we achieve an enlargement of the triple size of 100 % in the best case.

math.GR↗
arXiv2011

Methods of Matrix Multiplication: An Overview of Several Methods and their Implementation

In this overview article we present several methods for multiplying matrices and the implementation of these methods in C. Also a little test program is given to compare their running time and the numerical stability. The methods are: naive method, naive method working on arrays, naive method with the \textsc{Kahan} trick, three methods with loop unrolling, winograd method and the scaled variant, original \textsc{Strassen} method and the \textsc{Strassen}-\textsc{Winograd} variant. Please note, that this is the FIRST version. The algorithms are not well tested and the implementation is not optimized. If you like to join the project, please contact me.

cs.MS↗
arXiv2011

Strassen's Matrix Multiplication Algorithm for Matrices of Arbitrary Order

The well known algorithm of Volker Strassen for matrix multiplication can only be used for $(m2^k \times m2^k)$ matrices. For arbitrary $(n \times n)$ matrices one has to add zero rows and columns to the given matrices to use Strassen's algorithm. Strassen gave a strategy of how to set $m$ and $k$ for arbitrary $n$ to ensure $n\leq m2^k$. In this paper we study the number $d$ of additional zero rows and columns and the influence on the number of flops used by the algorithm in the worst case ($d=n/16$), best case ($d=1$) and in the average case ($d\approx n/48$). The aim of this work is to give a detailed analysis of the number of additional zero rows and columns and the additional work caused by Strassen's bad parameters. Strassen used the parameters $m$ and $k$ to show that his matrix multiplication algorithm needs less than $4.7n^{\log_2 7}$ flops. We can show in this paper, that these parameters cause an additional work of approx. 20 % in the worst case in comparison to the optimal strategy for the worst case. This is the main reason for the search for better parameters.

math.NA↗
arXiv2011

Using the Sherman-Morrison-Woodbury Formula to Solve the System of Linear Equations from the Standard Multiple Shooting Method for a Linear Two Point Boundary-Value Problem is a Bad Idea

We use the standard multiple shooting method to solve a linear two point boundary-value problem. To ensure that the solution obtained by combining the partial solutions is continuous and satisfies the boundary conditions, we have to solve a system of linear equations. Our idea is to first solve a bidiagonal system related to the original system of linear equations, and then update it with the Sherman-Morrison-Woodbury formula. We study the feasibility, the numerical stability and the running time of this method. The results are: The method described above has the same stability problems like the well known Condensing method. The running time analysis shows that the new method is slower than the Condensing method. Therefore we recommend not to use the method described in this article.

math.NA↗
arXiv2011

A Note on the Group-theoretic Approach to Fast Matrix Multiplication

In 2003 COHN and UMANS introduced a group-theoretic approach to fast matrix multiplication. This involves finding large subsets S, T and U of a group G satisfying the Triple Product Property (TPP) as a means to bound the exponent $ω$ of the matrix multiplication. We show that S, T and U may be be assumed to contain the identity and be otherwise disjoint. We also give a much shorter proof of the upper bound |S|+|T|+|U| <= |G|+2.

math.GR↗
arXiv2011

Search and test algorithms for Triple Product Property triples

In 2003 COHN and UMANS introduced a group-theoretic approach to fast matrix multiplication. This involves finding large subsets of a group $G$ satisfying the Triple Product Property (TPP) as a means to bound the exponent $ω$ of matrix multiplication. We present two new characterizations of the TPP, which are useful for theoretical considerations and for TPP test algorithms. With this we describe all known TPP tests and implement them in GAP algorithms. We also compare their runtime. Furthermore we show that the search for subgroup TPP triples of nontrivial size in a nonabelian group can be restricted to the set of all nonnormal subgroups of that group. Finally we describe brute-force search algorithms for maximal subgroup and subset TPP triples. In addition we present the results of the subset brute-force search for all groups of order less than 25 and selected results of the subgroup brute-force search for 2-groups, $SL(n,q)$ and $PSL(2,q)$.

math.GR↗