-
Krylov Solvers for Interior Point Methods with Applications in Radiation Therapy and Support Vector Machines
Authors:
Felix Liu,
Albin Fredriksson,
Stefano Markidis
Abstract:
Interior point methods are widely used for different types of mathematical optimization problems. Many implementations of interior point methods in use today rely on direct linear solvers to solve systems of equations in each iteration. The need to solve ever larger optimization problems more efficiently and the rise of hardware accelerators for general purpose computing has led to a large interes…
▽ More
Interior point methods are widely used for different types of mathematical optimization problems. Many implementations of interior point methods in use today rely on direct linear solvers to solve systems of equations in each iteration. The need to solve ever larger optimization problems more efficiently and the rise of hardware accelerators for general purpose computing has led to a large interest in using iterative linear solvers instead, with the major issue being inevitable ill-conditioning of the linear systems arising as the optimization progresses. We investigate the use of Krylov solvers for interior point methods in solving optimization problems from radiation therapy and support vector machines. We implement a prototype interior point method using a so called doubly augmented formulation of the Karush-Kuhn-Tucker linear system of equations, originally proposed by Forsgren and Gill, and evaluate its performance on real optimization problems from radiation therapy and support vector machines. Crucially, our implementation uses a preconditioned conjugate gradient method with Jacobi preconditioning internally. Our measurements of the conditioning of the linear systems indicate that the Jacobi preconditioner improves the conditioning of the systems to a degree that they can be solved iteratively, but there is room for further improvement in that regard. Furthermore, profiling of our prototype code shows that it is suitable for GPU acceleration, which may further improve its performance in practice. Overall, our results indicate that our method can find solutions of acceptable accuracy in reasonable time, even with a simple Jacobi preconditioner.
△ Less
Submitted 26 February, 2024; v1 submitted 1 August, 2023;
originally announced August 2023.
-
Parallel Cholesky Factorization for Banded Matrices using OpenMP Tasks
Authors:
Felix Liu,
Albin Fredriksson,
Stefano Markidis
Abstract:
Cholesky factorization is a widely used method for solving linear systems involving symmetric, positive-definite matrices, and can be an attractive choice in applications where a high degree of numerical stability is needed. One such application is numerical optimization, where direct methods for solving linear systems are widely used and often a significant performance bottleneck. An example wher…
▽ More
Cholesky factorization is a widely used method for solving linear systems involving symmetric, positive-definite matrices, and can be an attractive choice in applications where a high degree of numerical stability is needed. One such application is numerical optimization, where direct methods for solving linear systems are widely used and often a significant performance bottleneck. An example where this is the case, and the specific type of optimization problem motivating this work, is radiation therapy treatment planning, where numerical optimization is used to create individual treatment plans for patients. To address this bottleneck, we propose a task-based multi-threaded method for Cholesky factorization of banded matrices with medium-sized bands. We implement our algorithm using OpenMP tasks and compare our performance with state-of-the-art libraries such as Intel MKL. Our performance measurements show a performance that is on par or better than Intel MKL (up to ~26%) for a wide range of matrix bandwidths on two different Intel CPU systems.
△ Less
Submitted 8 May, 2023;
originally announced May 2023.
-
Digital Twins for Ports: Derived from Smart City and Supply Chain Twinning Experience
Authors:
Robert Klar,
Anna Fredriksson,
Vangelis Angelakis
Abstract:
Ports are striving for innovative technological solutions to cope with the ever-increasing growth of transport, while at the same time improving their environmental footprint. An emerging technology that has the potential to substantially increase the efficiency of the multifaceted and interconnected port processes is the digital twin. Although digital twins have been successfully integrated in ma…
▽ More
Ports are striving for innovative technological solutions to cope with the ever-increasing growth of transport, while at the same time improving their environmental footprint. An emerging technology that has the potential to substantially increase the efficiency of the multifaceted and interconnected port processes is the digital twin. Although digital twins have been successfully integrated in many industries, there is still a lack of cross-domain understanding of what constitutes a digital twin. Furthermore, the implementation of the digital twin in complex systems such as the port is still in its infancy. This paper attempts to fill this research gap by conducting an extensive cross-domain literature review of what constitutes a digital twin, kee** in mind the extent to which the respective findings can be applied to the port. It turns out that the digital twin of the port is most comparable to complex systems such as smart cities and supply chains, both in terms of its functional relevance as well as in terms of its requirements and characteristics. The conducted literature review, considering the different port processes and port characteristics, results in the identification of three core requirements of a digital port twin, which are described in detail. These include situational awareness, comprehensive data analytics capabilities for intelligent decision making, and the provision of an interface to promote multi-stakeholder governance and collaboration. Finally, specific operational scenarios are proposed on how the port's digital twin can contribute to energy savings by improving the use of port resources, facilities and operations.
△ Less
Submitted 21 August, 2023; v1 submitted 10 January, 2023;
originally announced January 2023.
-
Assessing the Maturity of Digital Twinning Solutions for Ports
Authors:
Robert Klar,
Anna Fredriksson,
Vangelis Angelakis
Abstract:
Ports are striving for innovative technological solutions to cope with the increasing growth in demand of goods transport, while at the same time improving their environmental footprint. An emerging technology that has the potential to substantially increase the effectiveness of the multifaceted and interconnected port processes is that of digital twins. Innovation-leading ports recognizing the po…
▽ More
Ports are striving for innovative technological solutions to cope with the increasing growth in demand of goods transport, while at the same time improving their environmental footprint. An emerging technology that has the potential to substantially increase the effectiveness of the multifaceted and interconnected port processes is that of digital twins. Innovation-leading ports recognizing the potential of twinning have already started working on it. However, since there is no clear consensus on what a digital twin of a complex system comprises and how it should be designed, deployed digital twin solutions for ports often differ significantly. This article addresses this issue by initially identifying three core aspect underpinning digital twins of complex systems, such as ports, and outlining five successive maturity levels based on these aspects' instantiation. These identified aspects and the derived maturity levels are then used to examine real-world cases by critically evaluating existing digital twinning solutions in the port of Singapore, the Mawan port of Shanghai, and that of Rotterdam. These being three of the world's innovation-leading ports, we naturally find in them most of the identified core aspects to be in line with their twinning implementation, which has reached, in all three, a higher level of maturity. Although, our work on maturity levels and core aspects can provide a guideline for designing and benchmarking future digital twinning solutions for ports, the capacity for innovation via twinning, even in the port domain, is highly contextual with key paragon being the availability of financial and technical resources.
△ Less
Submitted 15 December, 2022;
originally announced December 2022.
-
Distributed Objective Function Evaluation for Optimization of Radiation Therapy Treatment Plans
Authors:
Felix Liu,
Måns I. Andersson,
Albin Fredriksson,
Stefano Markidis
Abstract:
The modern workflow for radiation therapy treatment planning involves mathematical optimization to determine optimal treatment machine parameters for each patient case. The optimization problems can be computationally expensive, requiring iterative optimization algorithms to solve. In this work, we investigate a method for distributing the calculation of objective functions and gradients for radia…
▽ More
The modern workflow for radiation therapy treatment planning involves mathematical optimization to determine optimal treatment machine parameters for each patient case. The optimization problems can be computationally expensive, requiring iterative optimization algorithms to solve. In this work, we investigate a method for distributing the calculation of objective functions and gradients for radiation therapy optimization problems across computational nodes. We test our approach on the TROTS dataset -- which consists of optimization problems from real clinical patient cases -- using the IPOPT optimization solver in a leader/follower type approach for parallelization. We show that our approach can utilize multiple computational nodes efficiently, with a speedup of approximately 2-3.5 times compared to the serial version.
△ Less
Submitted 24 August, 2022;
originally announced August 2022.
-
Accelerating Radiation Therapy Dose Calculation with Nvidia GPUs
Authors:
Felix Liu,
Niclas Jansson,
Artur Podobas,
Albin Fredriksson,
Stefano Markidis
Abstract:
Radiation Treatment Planning (RTP) is the process of planning the appropriate external beam radiotherapy to combat cancer in human patients. RTP is a complex and compute-intensive task, which often takes a long time (several hours) to compute. Reducing this time allows for higher productivity at clinics and more sophisticated treatment planning, which can materialize in better treatments. The stat…
▽ More
Radiation Treatment Planning (RTP) is the process of planning the appropriate external beam radiotherapy to combat cancer in human patients. RTP is a complex and compute-intensive task, which often takes a long time (several hours) to compute. Reducing this time allows for higher productivity at clinics and more sophisticated treatment planning, which can materialize in better treatments. The state-of-the-art in medical facilities uses general-purpose processors (CPUs) to perform many steps in the RTP process. In this paper, we explore the use of accelerators to reduce RTP calculating time. We focus on the step that calculates the dose using the Graphics Processing Unit (GPU), which we believe is an excellent candidate for this computation type. Next, we create a highly optimized implementation for a custom Sparse Matrix-Vector Multiplication (SpMV) that operates on numerical formats unavailable in state-of-the-art SpMV libraries (e.g., Ginkgo and cuSPARSE). We show that our implementation is several times faster than the baseline (up-to 4x) and has a higher operational intensity than similar (but different) versions such as Ginkgo and cuSPARSE.
△ Less
Submitted 19 September, 2021; v1 submitted 17 March, 2021;
originally announced March 2021.
