Antonín Komenda

Czech Technical University in Prague | ČVUT · Department of Computer Science and Engineering (FEL)

Achieving joint objectives in distributed domain-independent planning problems by teams of cooperative agents requires significant coordination and communication efforts. For systems facing a plan failure in a dynamic environment, arguably, attempts to repair the failed plan in general, and especially in the worst-case scenarios, do not straightfor...

The development process and simulation architecture presented here help narrow the gap between how theoretical AI algorithms are traditionally designed and validated and how practical algorithms for controlling robotic assets in simulated tactical missions are developed.

Achieving joint objectives by teams of cooperative planning agents requires significant coordination and communication efforts. For a single-agent system facing a plan failure in a dynamic environment, arguably, attempts to repair the failed plan in general do not straightforwardly bring any benefit in terms of time complexity. However, in multi-ag...

Learning a well-informed heuristic function for hard planning domains is an elusive problem. Although there are known neural network architectures to represent such heuristic knowledge, it is not obvious what concrete information is learned and whether techniques aimed at understanding the structure help in improving the quality of the heuristics....

Automated planning provides a powerful general problem solving tool, however, its need for a model creates a bottleneck that can be an obstacle to using automated planning algorithms in real-world settings. In this work, we propose to use cellular simultaneous recurrent networks (CSRN), to process a planning problem and provide a heuristic value es...

Optimization of heuristic functions for the A* algorithm, realized by deep neural networks, is usually done by minimizing square root loss of estimate of the cost to goal values. This paper argues that this does not necessarily lead to a faster search of A* algorithm since its execution relies on relative values instead of absolute ones. As a mitig...

Privacy preservation has become one of the crucial research topics in multi-agent planning. A number of techniques to preserve private information throughout the planning process have emerged. One major difficulty of such research is the comparison of properties related to privacy among such techniques. A metric allowing for comparison of such priv...

Learning a well-informed heuristic function for hard task planning domains is an elusive problem. Although there are known neural network architectures to represent such heuristic knowledge, it is not obvious what concrete information is learned and whether techniques aimed at understanding the structure help in improving the quality of the heurist...

Automated planning for problems without an explicit model is an elusive research challenge. However, if tackled, it could provide a general approach to problems in real-world unstructured environments. There are currently two strong research directions in the area of artificial intelligence (AI), namely machine learning and symbolic AI. The former...

Privacy-Preserving Multi-Agent Planning (PP-MAP) has recently gained the attention of the research community, resulting in a number of PP-MAP planners and theoretical works. Many such planners lack strong theoretical guarantees, thus in order to compare their abilities w.r.t. privacy, a versatile and practical metric is crucial. In this work, we pr...

Potential heuristics assign a numerical value (potential) to each fact and compute the heuristic value for a given state as the sum of these potentials. A mutex is an invariant stating that a certain combination of facts cannot be part of any reachable state. In this paper, we use mutexes to improve potential heuristics in two ways. First, we show...

Multi-agent planning (MAP) has recently gained traction in both planning and multi-agent system communities, especially with the focus on privacy-preserving multi-agent planning, where multiple agents plan for a common goal but with private information they do not want to disclose. Heuristic search is the dominant technique used in MAP and therefor...

With the increasing prevalence of electric vehicles (EVs), the provision of EV charging is becoming a standard commercial service. With this shift, EV charging service providers are looking for ways to make their business more profitable. Dynamic pricing is a proven technique to increase revenue in markets with time-variant, heterogeneous demand. I...

Mutex groups are defined in the context of STRIPS planning as sets of facts out of which, maximally, one can be true in any state reachable from the initial state. This work provides a complexity analysis showing that inference of mutex groups is as hard as planning itself (PSPACE-Complete) and it also shows a tight relationship between mutex group...

Multi-agent planning can solve various sequential decision problems comprising multiple entities. In contrast to classical planning, the agents are interested in maintaining privacy while planning with each other. Therefore they have to reason about what information they can share. Although privacy is one of the crucial aspects of multi-agent plann...

Currently the most efficient distributed multiagent planning scheme for deterministic models is based on coordination of local agents’ plans. In such a scheme, behavior of other agents is modeled using projections of their actions stripped of all private information. The planning scheme does not require any additional information, however using suc...

Mutex groups are defined in the context of STRIPS planning as sets of facts out of which, maximally, one can be true in any state reachable from the initial state. The importance of computing and exploiting mutex groups was repeatedly pointed out in many studies. However, the theoretical analysis of mutex groups is sparse in current literature. Thi...

Multi-agent planning using MA-STRIPS–related models is often motivated by the preservation of private information. Such a motivation is not only natural for multi-agent systems but also is one of the main reasons multi-agent planning problems cannot be solved with a centralized approach. Although the motivation is common in the literature, the form...

Efficient allocation of charging capacity to electric vehicle (EV) users is a key prerequisite for large-scale adaption of electric vehicles. Dynamic pricing represents a flexible framework for balancing the supply and demand for limited resources. In this paper, we show how dynamic pricing can be employed for allocation of EV charging capacity. Ou...

Cooperative multi-agent planning (MAP) is a relatively recent research field that combines technologies, algorithms, and techniques developed by the Artificial Intelligence Planning and Multi-Agent Systems communities. While planning has been generally treated as a single-agent task, MAP generalizes this concept by considering multiple intelligent...

Real world applications often require cooperation of multiple independent entities. Classical planning is a well established technique solving various challenging problems such as logistic planning, factory process planning, military mission planning and high-level planning for robots. Multi-agent planning aims at solving similar problems in the pr...

Multi-agent planning using MA-STRIPS-related models is often motivated by the preservation of private information. Such motivation is not only natural for multi-agent systems, but it is one of the main reasons, why multi-agent planning (MAP) problems cannot be solved centrally. In this paper, we analyze privacy-preserving multi-agent planning (PP-M...

Unmanned aerial vehicles (UAVs) are suited to various remote sensing missions, such as measuring air quality. The conventional method of UAV control is by human operators. Such an approach is limited by the ability of cooperation among the operators controlling larger fleets of UAVs in a shared area. The remedy for this is to increase autonomy of t...

This article reports on the first international Competition of Distributed and Multiagent Planners (CoDMAP). The competition focused on cooperative domain-independent planners compatible with a minimal multiagent extension of the classical planning model. The motivations for the competition were manifold: to standardize the problem description lang...

Coordinated sequential decision making of a team of cooperative agents can be described by principles of multiagent planning. Provided that the mechanics of the environment the agents act in is described as a deterministic transitions system, an appropriate planning model is MA-Strips. Multiagent planning modeled as MA-Strips prescribes exactly wha...

Multi-agent planning using MA-STRIPS-related models is often motivated by the preservation of private information. Such motivation is not only natural for multi-agent systems, but is one of the main reasons, why multi-agent planning problems cannot be solved centrally. Although the motivation is common in the literature, formal treatment of privacy...

Distributed heuristic search is a well established technique for multi-agent planning. It has been shown that distributed heuristics may crucially improve the search guidance, but are costly in terms of communication and computation time. One solution is to compute a heuristic additively, in the sense that each agent can compute its part of the heu...

Similarly to classical planning, heuristics play a crucial role in most multi-agent and privacy-preserving multi-agent planning systems. It has been shown that distributed heuristics may crucially improve the search guidance, but are costly in terms of communication and computation time and are often a source of privacy concerns. One solution is to...

Multiagent planning addresses the problem of coordinated sequential decision making of a team of cooperative agents. One possible approach to multiagent planning, which proved to be very efficient in practice, is to find an acceptable public plan. The approach works in two stages. At first, a public plan acceptable to all the involved agents is com...

Distributed heuristic search is a well established technique for multi-agent planning. It has been shown that distributed heuristics may crucially improve the search guidance, but are costly in terms of communication and computation time. One solution is to compute a heuristic additively, in the sense that each agent can compute its part of the heu...

Reducing accidental complexity in planning problems is a well-established method for increasing efficiency of classical planning. Removal of superfluous facts and actions, and problem transformation by recursive macro actions are representatives of such methods working directly on input planning problems. Despite of its general applicability and th...

As a part of the workshop on Distributed and Multiagent Planning (DMAP) at the International Conference on Automated Planning and Scheduling (ICAPS) 2015, we have organized a competition in distributed and multiagent planning. The main aims of the competition were to consolidate the planners in terms of input format; to promote development of multi...

Agents planning under STRIPS-related model using separation of facts and actions to private and public can model behavior of other agents as public external projections of their actions. In the most simplistic case, the agent does not require any additional information from the other agents, that is the planning process ignores any dependencies of...

Deterministic domain-independent multiagent planning is an approach to coordination of cooperative agents with joint goals. Provided that the agents act in an uncertain and dynamic environment, such plans can fail. The straightforward approach to recover from such situations is to compute a new plan from scratch, that is to replan. Even though, in...

Heuristics are a crucial component in modern planning systems. In optimal multiagent planning the state of the art is to compute the heuristic locally using only information available to a single agent. This approach has a major deficiency as the local shortest path can arbitrarily underestimate the true shortest path cost in the global problem. As...

Interruptible pure exploration in multi-armed bandits (MABs) is a key component of Monte-Carlo tree search algorithms for sequential decision problems. We introduce Discriminative Bucketing (DB), a novel family of strategies for pure exploration in MABs, which allows for adapting recent advances in non-interruptible strategies to the interruptible...

Multiagent planning is a coordination technique used for deliberative acting of a team of agents. One of vital planning techniques uses declarative description of agents’ plans based on Finite State Machines and their later coordination by intersection of such machines with successive verification of the resulting joint plans. In this work, we firs...

Multiagent planning is a coordination technique used for deliberative acting of a team of agents. One of vital planning techniques uses declarative description of agents' plans based on Finite State Machines and their later coordination by intersection of such machines with successive verification of the resulting joint plans. In this work, we firs...

Heuristics are a crucial component in modem planning systems. In optimal multiagent planning the state of the art is to compute the heuristic locally using only information available to a single agent. This approach has a major deficiency as the local shortest path can arbitrarily underestimate the true shortest path cost in the global problem. As...

Coordinated sequential decision making of a team of cooperative agents is described by principles of multiagent planning. In this work, we extend the MA-Strips formalism with the notion of extensibility and reuse a well-known initiator–participants scheme for agent negotiation. A multiagent extension of the Generate-And-Test principle is used to di...

Deterministic multi-agent planning described by MA-STRIPS formalism requires mixture of coordination and synthesis of local agents' plans. All agents' plans, as sequences of actions, can be implicitly described by an appropriate generative structure. Having all local plans of all participating agents described by such a structure and having a merge...

In online planning with a team of cooperative agents, a straightforward model for decision making which actions the agents should execute can be represented as the problem of Combinatorial Multi-Armed Bandit. Similarly to the most prominent approaches for online planning with polynomial number of possible actions, state-of-the-art algorithms for on...

Similarly to classical planning, in MA-STRIPS multiagent planning, heuristics significantly improve efficiency of search-based planners. Heuristics based on solving a relaxation of the original planning problem are intensively studied and well understood. In particular, frequently used is the delete relaxation, where all delete effects of actions a...

Similarly to classical planning, in MA-Strips multiagent planning, heuristics significantly improve efficiency of search-based planners. Heuristics based on solving a relaxation of the original planning problem are intensively studied and well understood. In particular, frequently used is the delete relaxation, where all delete effects of actions a...

Multiagent planning for cooperative agents in deterministic environments intertwines synthesis and coordination of the local plans of involved agents. Both of these processes require an underlying structure to describe synchronization of the plans. A distributed planning graph can act as such a structure, benefiting by its compact representation an...

Problems of domain-independent multiagent planning for cooperative agents in deterministic environments can be tackled by a well-known initiator–participants scheme from classical multiagent negotiation protocols. In this work, we use the approach to describe a multiagent extension of the Generate-And-Test principle distributively searching for a c...

Online planning algorithms are typically a tool of choice for dealing with sequential decision problems in combinatorial search spaces. Many such problems, however, also exhibit combinatorial actions, yet standard planning algorithms do not cope well with this type of 'the curse of dimensionality'. Following a recently opened line of related work o...

Achieving joint objectives in distributed domain-independent planning problems by teams of co- operative agents requires significant coordination and communication efforts. Provided that the agents act in an imperfect environment, their plans can fail. The straightforward approach to re- cover from such situations is to compute a new plan from scra...

Small Unmanned Aerial Vehicles (UAVs) are becoming increasingly popular for tasks such as surveillance or target tracking in various types of tactical missions. Traditionally, each UAV in a mission is controlled by one or more operators. In our previous work we have developed a collection of distributed algorithms that allow one operator to control...

Deterministic domain-independent planning techniques for multiagent systems stem from principles of classical planning. Three most recently studied approaches comprise (i) DisCSP+Planning utilizing Distributed Constraint Satisfaction Problem solving for coordination of the agents and individual planning using local search, (ii) multiagent adaptatio...

Since late 90's of the last century, rapid advances in technology, mechanical engineering, miniaturization, telecommunications and informatics enabled development and routine deployment of sophisticated robots in many real world domains. Besides many applications in assembly industry, e.g., in car, or electronics assembly lines, defense organizatio...

AgentPolis is a fully agent-based platform for modeling multi-modal transportation systems. It comprises a high-performance discrete-event simulation core, a cohesive set of high-level abstractions for building extensible agent-based models and a library of predefined components frequently used in transportation and mobility models. Together with a...

With scaling of multi-robot teams deployed in military operations, there is a need to boost autonomy of individual, as well as team behaviors. We developed a feature-rich simulation testbed for experimental evaluation of multi-agent coordination mechanisms applicable in tactical military operations in urban warfare. In particular, we investigated a...

This paper proposes a mechanism for simulating limited communication bandwidth and processing power available to an agent in multi-agent simulations. Although there exist dedicated tools able to simulate computer networks, most multi-agent platforms lack support for this kind of resource allocation. We target such multi-agent platforms and offer an...

Autonomous control of group of unmanned aerial vehicles based on task allocation mechanisms shows great potential for ground tactical mission support. We introduce experimental simulation system combining flexible mission control of ground assets in urban environment and autonomous aerial support utilizing multi-agent problem solving techniques. Tw...

We present an approach to plan representation in multi-actor scenarios that is suitable for flexible replanning and plan revision purposes in dynamic non-deterministic multi-actor environments. The key idea of the presented approach is in representation of the distributed hierarchical plan by social commitments, as a theoretically studied formalism...

Problem solving and planning in decentralized environments is a key technical challenge in numerous industrial applications, ranging from manufacturing, logistics, virtual enterprizes to multirobotics systems. We present an abstract architecture of a multiagent solver and respective algorithm providing decomposition, task allocation, and task deleg...

Considering Unmanned Autonomous Vehicles (UAVs) the planning tasks mainly consist of finding paths between given waypoints with respect to given constraints. In this paper we developed a path planning system for flying UAVs (VTOLs and CTOLs) built upon Hexagonal grids which also supports simple dynamics (handling with speed). The planning system is...

This paper proposes a mechanism for simulating limited communication bandwidth and processing power available to an agent in multi-agent simulations. Although there exist dedicated tools able to simulate computer networks, most multi-agent platforms lack support for this kind of resource allocation. We target such multi-agent platforms and offer an...

Navigation of unmanned ground vehicles in an urban area is a fundamental problem which has to be solved prior to real-world deployment of the autonomous ground assets. Since the topology data of the environment are usually known a priori, they can be exploited in high-level planning of the routes. On the other hand, the low-level robot control requ...

A multi-agent VRP solver is presented in this paper. It utilizes the contract-net protocol based allocation and several improvement strategies. It provides the solution with the quality of 81% compared to the optimal solution on 115 benchmark instances in polynomial time. The self-organizing capability of the system successfully minimizes the numbe...