Glosarium AI
Kamus lengkap Kecerdasan Buatan
Distributed Multi-Agent Planning
Approach where multiple autonomous agents compute their trajectories in a decentralized manner, without a central supervisor, based on local information and inter-agent communications to achieve a global objective.
Distributed Constraint Satisfaction Problem (DCSP)
Mathematical formalization where each agent manages a subset of variables and constraints, and the goal is to find an assignment of values to all variables that satisfies the set of inter-agent constraints.
Distributed Consensus Search
Iterative process by which agents exchange information about their local states or decisions to converge towards a common agreement or solution, essential for action synchronization.
Coupled Trajectory Optimization
Technique where the trajectories of multiple agents are optimized jointly, taking into account their dynamic interdependencies, to avoid collisions and improve the overall system efficiency.
Distributed Potential Cost Function
Method where each agent defines a cost function that includes potential terms to model the influence of other agents, enabling decentralized optimization that integrates interactions.
Distributed Receding Horizon Planning (Distributed Model Predictive Control)
Control strategy where each agent solves an optimization problem over a sliding horizon, anticipating the future actions of other agents and adjusting its own trajectory accordingly.
Temporal Dependency Graphs
Data structures representing precedence and temporal synchronization constraints between the actions of different agents, used to ensure ordered and coherent plan execution.
Distributed Task and Resource Allocation
Mechanism by which agents negotiate or coordinate to distribute objectives and spatial or temporal resources among themselves, in order to maximize collective efficiency while minimizing conflicts.
Event-Based Synchronization
Coordination paradigm where agents adjust their progress based on triggering events (e.g., reaching a waypoint, detecting another agent), rather than on globally synchronized time.
Collision-Free Multi-Agent Routing
Set of algorithms ensuring that paths computed for each agent in a shared environment are free from any spatial and temporal intersection that could lead to physical collision.
Distributed Simultaneous Localization and Mapping (Distributed SLAM)
Process where a team of agents collaborates to build a common map of an environment while estimating their own trajectories, by fusing sensory data in a decentralized manner.
Constrained Multi-Agent Formation
Planning problem where agents must maintain a specific relative geometry (a formation) while collectively navigating toward a goal, requiring continuous coordination of speeds and positions.
Distributed Temporal Logic
Extension of temporal logics (such as LTL) to specify and verify properties on the global behavior of a multi-agent system, where logical operators are evaluated on distributed execution traces.
Trajectory Negotiation
Communication process by which agents propose, counter-propose, and accept modifications to their initial trajectories to resolve intersection conflicts and reach a mutually acceptable plan.
Distributed Robust Planning Under Uncertainty
Approach aiming to generate trajectories for each agent that remain valid and safe despite model inaccuracies, communication delays, or unpredictable actions of other agents.
Conflict Point Passage Scheduling
Coordination mechanism that assigns time slots for agents to pass through critical zones (intersections, narrow passages), in order to sequence their traversal and ensure smooth flow without deadlock.
Multi-Agent Reinforcement Learning for Planning
Use of learning algorithms where each agent learns a planning policy (trajectory choice) by interacting with the environment and other agents, to maximize a collective or individual reward.