The goal of this workshop is to create a forum for scientists, engineers and practitioners to present, discuss and collaborate on new results/ideas in the intersection of algorithms research, computational sciences, decision sciences and optimization. There is a growing recognition that a variety of use-cases in modern science and engineering, including social sciences and engineering applications, are poised to benefit from fast real-time decision-making capabilities. Increasingly powerful computing technologies are expected to solve optimization and decision science problems that were considered computationally prohibitive for integration within scientific and engineering workflows, especially those needing real-time decision-making capabilities. The workshop encourages topics on the theory, design and applications of novel methods and algorithms for different types of optimization problems that arise within the broader purview of decision-making capabilities in areas such as scientific/engineering workflows, logistics, transportation and urban planning, public health, manufacturing, energy (e.g., electric grids), digital twin systems (e.g., precision agriculture, smart cities, earth systems), operations management, machine learning, finance and others. Both regular papers as well as short position papers describing work-in-progress with innovative ideas related to the workshop topics are being solicited.

Important Dates

For information on PDADS 2024, please visit: https://www.csm.ornl.gov/workshops/PDADS2024/index.html .

IMPORTANT DATES

• Paper Submission Deadline: June 2, 2024 (AoE) [Hard Deadline]
• Author Notification: June 16, 2024 (AoE)
• Camera-Ready Deadline: June 26, 2024 (AoE)
• Workshop: August 12, 2024

Submission Guidelines

This workshop will focus on research at the intersection of parallel and distributed algorithms, decision sciences and combinatorial optimization. The workshop will discuss latest trends and identify technology gaps in high-performance decision sciences and combinatorial optimization technologies for extant and next-generation scientific, engineering and other applications. The workshop adopts an inclusive definition of the sciences that includes the social sciences, behavioral sciences or others. Both regular papers and short position papers describing work-in-progress with innovative ideas related to the workshop topics are being solicited. Accepted papers will be published by ACM ICPS in a workshop proceedings volume to be made available for download from the ACM digital library. Topics of interest include, but are not limited to: 

• Novel AI applications in systems and decision sciences on parallel computing systems. 
• Deep learning solutions to optimization problems, such as reinforcement learning for control, combinatorial optimization, and decision making. 
• Generative AI approaches to scenario analysis for decision-making support 
• Scalable data driven decision-making methods and algorithms powered by machine learning.
• Deep learning model deployment and performance evaluation in decision support environments. 
• Decision support foundation models that leverage large language models (LLMs) or are trained from scratch.
• Optimization techniques in machine learning, such as high-performance first and higher order iterative optimization algorithms for minimizing loss and optimizing weight and bias tensors. 
• Application-centric manuscripts involving optimizations for decision-making capabilities in systems such as logistics, transportation and urban planning, public health, manufacturing, energy (e.g., electric grids), digital twin systems (e.g., precision agriculture, smart cities, earth systems) operations management, finance and other areas are especially encouraged.
• High-performance algorithms for integer/mixed-integer programming, linear/nonlinear programming, stochastic programming, robust optimization, combinatorial optimization, feasibility problems (SAT, CP, etc.). 
• High-performance heuristic and meta-heuristic algorithms. 
• High-performance local and complete search methods. 
• Learning approaches for optimization in parallel and distributed environments. 
• Parallel and distributed approaches for parameter tuning, simulation-based optimization, and black box optimization. 
• Parallel algorithm portfolios. 
• Quantum optimization algorithms. 
• Use of randomization techniques for scalable decision support systems. 
• Application of decision support systems on novel computing platforms (shared/distributed memory, edge devices, cloud platforms, field programable gate arrays, GPU, TPU, quantum computers, etc.). 
• Use of parallel and distributed computing for timely and/or higher quality decision support. 
• Theoretical analysis of convergence and/or complexity of parallel optimization algorithms and decision support systems.
• High-performance deep learning solutions that harness various parallelisms in deep learning computing for scalable acceleration on GPU/TPU.

Committees

Program Committee

• TBD

Organizing committee

• Sudip K. Seal, Oak Ridge National Laboratory
• Yan Liu, Oak Ridge National Laboratory
• Daniel Pack, Oak Ridge National Laboratory

Publication

The Fourth PDADS-2024 Workshop Proceedings will be published as a separate volume along with the ICPP 2024 conference proceedings. More details coming soon!

Venue

The conference will be held in Gotland, Sweden and co-hosted with the 53rd International Conference on Parallel Processing (ICPP) between August 12-15, 2024.

Contact

All questions about submissions should be emailed to sealsk@ornl.gov or yanliu@ornl.gov.