- Artificial intelligence:
automated planning and scheduling, constraint satisfaction, knowledge representation, robotics, machine learning
- At SOFSEM 2002 paper, I proposed the constraint for modeling sequences of states based on finite-state automaton (and the corresponding filtering algorithm). This constraint is now known as the regular or automaton constraint and it is used in most constraint solvers.
- At AOR 2003 paper, I proposed the concept of dynamic global constraints (currently known as open global constraints). These constraints allow dynamic extension of the set of variables, which is motived by planning problems.
- With Tomáš Müller and Hana Rudová we formalised the concept of minimal perturbation problem describing the objective to minimize the number of changes in the problem solution after problem modification (PATAT 2004).
- With Petr Vilím and Ondųej Čepek we proposed efficient (n.log n) filtering algorithms for unary resource constraints with optional activities. This significantly extends modeling capabilities of constraint-based scheduling to cover alternative procesess (planning). It is now used in systems such as IBM CP Optimizer and Choco. The paper got the Best Paper Award at CP 2004.
- Together with Ondųej Čepek we proposed the concept of Temporal Networks with Alternatives that extends simple temporal networks to cover alternative branches (FLAIRS 2007).
- With Daniel Toropila, we proposed several constraint models for classical planning problems (FLAIRS 2008, CP-AI-OR 2009, SCAI 2011). It may be worth to investigate how these models work with modern CP/SAT solvers.
- I led the team that developed the FlowOpt scheduling system based on concept of temporal networks with alternatives. The system covers process editor, scheduler, interactive re-scheduler, and schedule analyser (ECAI 2012) and some ideas are used in the software MAKE by Entellexi, Ltd. The FlowOpt system got the Best System Demonstration Award at ICAPS 2012.
- I advocate the importance of formal models in efficiency of problem solving and planning in particular. In particular, together with Neng-Fa Zhou, Agostino Dovier, and Lukáš Chrpa we showed that including even simple common sense knowledge in the model can beat the most sophisticated automated planners (TPLP 2014, PPDP 2015, TPLP 2015, KR 2016).
- Togeter with Adrien Maillard (and Rafael C. Cardoso) we suggested attribute grammars as a unifying framework for modeling planning and scheduling problems. We proposed the first formal verification algorithm for full hierarchical temporal networks (ICAPS 2018).
- With my PhD students, Jiųí Švancara and Marek Vlk, and in co-operation with the team at Ben-Gurion University (led by Ariel Felner and Roni Stern) and Neng-Fa Zhou (Brooklyn College), I work on the problem of multi-agent path finding, in particular on compilation approaches to solve the problem. I advocate bringing the academic problem closer to real-life by proposing more relatistic formal models (AAMAS 2018) and testing the results on real robots (IBERAMIA 2018, Best Paper Award).
Visopt Ltd. (Israel), Entellexi (Ireland), NASA (USA), ESA (Germany), Universal Synergetics (USA), AISA/Median (Czech Republic)
Masaryk University (Czech Republic), Czech Technical University (Czech Republic), Czech Academy of Sciences (Czech Republic), University of Limerick (Ireland), Cork Constraint Computation Centre/University College Cork (Ireland), Aalborg University (Denmark), Ben-Gurion University (Israel), Brooklyn College CUNI (United States), University of Southern California (United States)
Publications and metrics:
- Hierarchical Planning: From Plan Verification to Plan Recognition (CSF P202-21-13882J), 2021-2023
- Lifelong Planning for a Smart Swarm (M©MT LTAUSA19072), 2020-2023
- Diagnosis and Troubleshooting for the Execution of Multi-Robot-Path-Finding Plans (MŠMT LTAIZ19014), 2019-2022
- Smart Swarms: From Theory to Practice (CSF P103-19-02183S), 2019-2022
- MoRePlan: Modeling and Reformulating Planning Problems (CSF P103-18-07252S), 2018-2021
- Integration Of Heuristic Search And Compilation-Based Techniques For Multi-Agent Path-Finding (MŠMT 8G15027), 2016-2018
- Automated Knowledge and Plan Modeling for Autonomous Robots (CSF P103-15-19877S), 2015-2017
- PlanEx: Bridging Planning and Execution (CSF P103/10/1287), 2010-2014
- KnowSched: Knowledge Techniques in Scheduling (CSF P202/10/1188), 2010-2013)
- Res Informatica (CSF 201-09-H057), 2009-2012
- ValuePOLE – An Extended Value Chain Model for Performnace Prediction and Optimisation of Product and Process Lifecycles for SMEs (FP7 222218), 2008-2011
- LeCoS: merging machine LEarning and COnstraint Satisfaction (CSF 201/08/0509), 2008-2010
- Dynamic Aspects of Scheduling (CSF 201/07/0205), 2007-2009
- Collegium Informaticum (CSF 201/05/H014), 2005-2008
- EMPOSME-Enterprise Modelling and Performance Optimisation (FP6), 2005-2008
- Planning and scheduling with constraints (CSF 201/04/1102), 2004-2006
- Advanced Planning and Scheduling (CSF 201/01/0942), 2001-2003
- Logic Programs with Constraints (CSF 201/99/D057), 1999-2001
If you are interested to run a joint project with my team, please contact me.