Jiří Vokřínek

Competitive Contract Net Protocol

The proposed Competitive Contract Net Protocol has been designed to facilitate a flexible cooperation in competitive multi-agent environments and to support automated or semi-automated negotiations in competitive domains. The protocol is based on FIPA standards. The protocol covers not only the phase of contracting the commitments, but also allows for a decommitment negotiation and contract termination. Thus, it consists of three phases: (i) a contracting phase, where conditions of agreement are concluded, (ii) an optional decommitment phase, where contract may be breached, and (iii) a contract termination phase, where the compliance with the concluded contract conditions is evaluated. Both the decommitment and non-compliance are bounded with penalties which measurably ensure a compliance with the commitments, but also allow an opportunistic behaviour of the agents at some price.

Agents Toward Vehicle Routing Problem With Time Windows

We suggest an efficient polynomial algorithm for the vehicle routing problem with time windows based on agent negotiation. A generic agent decomposition of the problem is introduced featuring a clear separation between the local planning performed by the individual vehicles and the abstract global coordination achieved by negotiation—differentiating the presented algorithm from the classical centralized algorithms. The semantic of the abstract negotiation process is discussed, as well as the two alternative local planning strategies used by the individual vehicles. A relevant comparison to the state-of-the-art classical and agent-based algorithms is provided, missing from most previous agent-based studies. The algorithm performs slightly worse than the state-of-the-art classical algorithms but improves on all the previous comparable agent-based algorithms.

Competencies and Profiles Management for Virtual Organizations Creation

Sharing information about profiles and offered competencies of individual members within an alliance of cooperating companies facilitates searching for potential members of Virtual Organizations. This paper presents a concept of structuring of the competencies as well as a prototype proving this concept. The prototype follows a naturally hybrid architecture of alliances that consists of mutually independent alliance partners, who may be supported by central institutions of the alliance. The use of the agent-based solution enables information sharing among partners in such distributed and dynamic environment.

Multi-Agent Traffic Simulation for Development and Validation of Autonomic Car-to-Car Systems

In this chapter, we present the concept of an integrated multi-agent simulation platform to support the development and validation of autonomic cooperative car-to-car systems. The simulation allows to validate the car-to-car coordination strategies in various traffic scenarios in variable technology penetration levels (i.e. mixing different strategies) and user acceptance of such system as an external observer and/or as a part of the traffic (human in the loop with intelligent cooperative guidance system). The platform combines features of realistic driving simulation, traffic simulation with flexible level of detail and AI controlled vehicles. The principal idea of the platform is to allow the development and study of complex autonomic distributed car-to-car systems for vehicles coordination. The platform provides a development environment and a tool chain that is necessary for the validation of such complex systems. Autonomic car-to-car systems are based on coordination mechanisms between agents, where an agent represents a reasoning unit of a single vehicle. The road traffic is modelled as a multi-agent system of cooperative agents. The interaction between the agents brings autonomic properties into the emerged system (e.g. the traffic adapts to a blockage of a lane and vehicles merge into a second lane). The system also exhibits autonomic properties from a single user perspective. The driver approaches the system in a form of a driver assistance system—we can refer it as an autonomic driver assistance system. The driver is interacting only with the assistance system via a human-machine interface (HMI). The autonomic driver assistance system is hiding the complexity of multi-agent interactions from the user. The related agent of the single vehicle is responsible for an interaction with other agents in the system without any user’s intervention.

The Art of Negotiation: Developing Efficient Agent-Based Algorithms for Solving Vehicle Routing Problem with Time Windows

We present an ongoing effort in developing efficient agent-based algorithms for solving the vehicle routing problem with time windows. An abstract algorithm based on a generic agent decomposition of the problem is introduced featuring a clear separation between the local planning performed by the individual vehicles and the global coordination achieved by negotiation. The semantics of the underlying negotiation process is discussed as well as the alternative local planning strategies used by the individual vehicles. Finally a parallel version of the algorithm is presented based on efficient search diversification and intensification strategies. The presented effort is relevant namely for i yielding results significantly improving on all previous agent-based studies, ii the inclusion of relevant widely-used benchmarks missing from these studies and iii the breadth and depth of the provided evidence and analysis including relevant comparison to the state-of-the-art centralized solvers.

Ground tactical mission support by multi-agent control of UAV operations

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. Two case-studies are presented for evaluation - cooperative area surveillance and dynamic target tracking with undervalued number of assets. We show the strength and benefits of multi-agent task allocation and delegation mechanisms in such dynamic scenarios mainly in case of limited number of assets.

Plan representation and execution in multi-actor scenarios by means of social commitments

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 representing mutual relations among intentions of collaborating agents. The article presents a formal model of a recursive form of commitments and discusses how it can be deployed to a selected hierarchical planning scenario. The decommitment rules definition and their influence on the plan execution robustness and stability is also presented. The approach was verified and evaluated in a simulated environment. The experimental validation confirms the performance, stability, and robustness of the system in complex scenarios.

Cooperative agent navigation in partially unknown urban environments

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 requires precise path to follow and thus trajectory planning has to be adopted as well. Finally, the particular details of the environment can differ from the known topology and thus the vehicles need an area exploration method. The integrated multi-agent system for cooperative navigation in partially unknown urban environment is introduced and validated using vehicle physics based simulation. The goal is to support the navigation of a convoy by a set of unmanned vehicles in an urban environment to avoid convoy stops or u-turns and minimize the total traveled distance of all vehicles. The high-level planning component (long-time horizon) is based on distributed problem solver (DPS) together with a D-star route planner, where the DPS solves a dynamic vehicle routing problem over a set of generated frontiers describing interesting points in the map. After allocation of the tasks, each vehicle uses an adaptive path planner (short-time horizon) for generation of the precise path, which is then followed. Any new information sensed by the vehicles is additionally integrated into the global knowledge and can influence the planning processes.

Simulated Multi-robot Tactical Missions in Urban Warfare

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 organizations, together with space exploration and mining industries belong to the most demanding and optimistic users of robotic technology [25]. Especially in the military domain we nowadays witness a routine deployment of robotic assets in the field.

AgentDrive: Agent-Based Simulator for Intelligent Cars and Its Application for Development of a Lane-Changing Assistant

Intelligent cars represent a promising technology expected to drastically improve safety and efficiency of automobile transportation. In this paper, we introduce an agent-based simulation platform AgentDrive and argue that it can be used to speed up the development and evaluation of new coordination algorithms for intelligent cars. We present the high-level architecture of the simulator and characterize the class of tasks for which is the tool best suited. In addition, we present a case study of AgentDrive being used for development of a lane-changing assistant technology. We describe the developed solution in detail and present the benchmark result, which were obtained using AgentDrive simulator, that demonstrate that coordinated lane changing enables safer and swifter lane changing then the traditional non-coordinated approach.