Abstract The papers in this proceedings present the latest advances in the field of automated planning and scheduling, ranging in scope from theoretical analyses of planning and scheduling problems and processes, to new algorithms for planning and scheduling under various constraints and assumptions, and the empirical evaluation of planning and scheduling techniques. They reflect recent research trends in subareas such as optimal planning, probabilistic and nondeterministic planning, path planning, multiagent planning, and new developments in heuristics and their analysis for planning algorithms.
Introduction
The 24th International Conference on Automated Planning and Scheduling (ICAPS 2014) serves as a comprehensive overview of the state of the field in 2014. As the premier forum for researchers and practitioners in automated planning and scheduling, the conference proceedings reflect a rigorous selection process and highlight emerging trends that define the current trajectory of artificial intelligence. This preface outlines the organizational structure, the thematic shifts toward practical application, and the community-driven initiatives that characterized the 2014 program.
Organizational Structure and Selection Process
Maintaining the high archival standards established by its predecessor conferences (AIPS and ECP), ICAPS 2014 employed a rigorous peer-review process. The conference achieved an overall acceptance rate of just under 38%, ensuring that the published work represents the highest quality research in the field.
To manage the expanding scope of the discipline, the conference was organized into several specialized tracks:
- Main Track: Received 112 submissions, with 39 long and 7 short papers accepted.
- Novel Applications Track: In its second year, this track received 24 submissions, accepting 6 long and 1 short paper. This track specifically addresses the gap between theoretical algorithms and industrial deployment.
- Robotics Track: Debuting in 2014, this track received 28 submissions, with 7 long and 2 short papers accepted. It marks a significant effort to integrate planning as the "brain" of robotic systems.
- Journal Presentation Track: Designed to bring deep, complex research previously published in journals back into the live conference environment, this track accepted 8 out of 19 submissions.
Key Research Trends and Insights
The 2014 proceedings highlight several critical shifts in the planning and scheduling landscape:
The "Robotics Turn"
The introduction of a dedicated Robotics track reflects a growing realization that automated planning is essential for real-world robotic execution. Key research areas included the combination of task and motion planning, human-robot interaction, and planning specifically for perception-heavy environments. This integration suggests that the future of the field lies in moving from abstract problem-solving to embodied intelligence.
Uncertainty and Heuristic Evolution
Research trends have increasingly moved away from "perfect world" assumptions. There is a noted emphasis on probabilistic and non-deterministic planning, addressing the inherent uncertainty of real-world environments. Furthermore, the development and analysis of new heuristics remain a cornerstone of the field, as researchers seek to make planning algorithms computationally feasible for increasingly complex, large-scale problems.
Interdisciplinary Convergence
The Journal Track and invited talks (featuring leaders like Malik Ghallab, Michael Littman, and Peter Wurman) underscored that planning is no longer an isolated sub-field. It increasingly intersects with constraint programming, machine learning, and operations research, creating a more holistic approach to intelligent decision-making.
Community and Educational Outreach
Beyond the presentation of technical papers, ICAPS 2014 emphasized the growth of the research community through several initiatives:
- Tutorials and Workshops: The program included 7 tutorials and 8 workshops covering diverse topics such as Relational Dynamic Decisional Language (RDDL) and multi-agent pathfinding. These sessions provided essential training for the next generation of researchers.
- Doctoral Consortium: A dedicated forum for PhD students ensured mentorship and visibility for emerging scholars.
- System Demonstrations: This track focused on the actual software implementations of planning systems, highlighting the community's commitment to moving from theory to functional tools.
Conclusion
ICAPS 2014 successfully curated a program that balances theoretical depth with empirical evaluation. While the conference celebrated significant milestones—particularly the successful debut of the Robotics track—it also acknowledged ongoing challenges. The path from a theoretical algorithm to a deployed real-world system remains a significant hurdle, and the integration of planning with robotics is viewed as a long-term project. These proceedings stand as a testament to a maturing field that is increasingly focused on the complexities of the real world.
Summary updated