In this paper, a cooperative driver model for a multi-agent traffic simulation is proposed. The model combines maneuver-based trajectory planning of the vehicles with a cooperative conflict resolving. The proposed model is able to provide a safe drive in complex traffic situations at the highest possible speed. The idea of the model and its feasibility have been verified in complex scenarios such as line change under heavy traffic, highway entering or highway crossing. Moreover, the developed cooperative driver model is being integrated with a human operated driving simulator that enables verification of the proposed model in mixed scenarios enriching the simulation for a human driver with highly cooperative background traffic; thus, providing a platform for further studies on benefits of assistive technologies. The paper provides description of the proposed model and its early evaluation on the selected scenarios in a multi-agent traffic simulation.
In coalition formation with self-interested agents both social welfare of the multi-agent system and stability of individual coalitions must be taken into account. However, in large-scale systems with thousands of agents, finding an optimal solution with respect to both metrics is infeasible.
Avoiding collisions is one of the vital tasks for systems of autonomous mobile agents. We focus on the problem of finding continuous coordinated paths for multiple mobile disc agents in a 2-d environment with polygonal obstacles. The problem is PSPACE-hard, with the state space growing exponentially in the number of agents. Therefore, the state of the art methods include mainly reactive techniques and sampling-based iterative algorithms. We compare the performance of a widely-used reactive method ORCA with three variants of a popular planning algorithm RRT* applied to multi-agent path planning and find that an algorithm combining reactive collision avoidance and RRT* planning, which we call ORCA-RRT* can be used to solve instances that are out of the reach of either of the techniques. We experimentally show that: 1) the reactive part of the algorithm can efficiently solve many multi-agent path finding problems involving large number of agents, for which RRT* algorithm is often unable to find a solution in limited time and 2) the planning component of the algorithm is able to solve many instances containing local minima, where reactive techniques typically fail.
Groundwater is one of the most vital of all common pool resources throughout the world. More than half of groundwater is used to grow crops. This research models groundwater depletion patterns within a multi-agent system framework. Irrigators are modeled as agents in the multi-agent system. The irrigation strategies adopted by the agents are investigated using game theory. A set of five irrigators, growing three crops: corn, sorghum and wheat, have been considered in this study. To allow groundwater flow, these agents are assumed to be located in adjoining farm lands. Irrigators are modeled selfish agents that strategize their irrigation patterns in order to maximize their own utilities, i.e. the difference between the total revenue obtained from crop sales and the costs incurred, including groundwater extraction costs. Due to groundwater flow, and have no incentive to conserve groundwater. This leads to unsustainable depletion of the resource under Nash equilibrium, when no irrigator can increase its utility by unilaterally changing its strategy. All parameters in this research are representative of Kansas. Recorded environmental and economic data of the region, along with the DSSAT software, have been used to obtain these futuristic projections. One of the emergent phenomena of the simulations is the adoption of crop rotation patterns by the irrigators to conserve groundwater. The irrigators grow corn, which is a more profitable yet water intensive crop in one year, and in the next, conserve water by growing sorghum instead. Another emergent outcome of this research is the viability of LEMAs. When the irrigators are subject to LEMA-level limits on groundwater use, there is a slight increase in the aggregate utility of the LEMA.
Avoiding collisions is one of the vital tasks for systems of autonomous mobile agents. We focus on the problem of finding continuous coordinated paths for multiple mobile disc agents in a 2-d environment with polygonal obstacles. The problem is PSPACE-hard, with the state space growing exponentially in the number of agents. Therefore, the state of the art methods include mainly reactive techniques and sampling-based iterative algorithms.We compare the performance of a widely-used reactive method ORCA with three variants of a popular planning algorithm RRT* applied to multi-agent path planning and find that an algorithm combining reactive collision avoidance and RRT* planning, which we call ORCA-RRT* can be used to solve instances that are out of the reach of either of the techniques. We experimentally show that: 1) the reactive part of the algorithm can efficiently solve many multi-agent path finding problems involving large number of agents, for which RRT* algorithm is often unable to find a solution in limited time and 2) the planning component of the algorithm is able to solve many instances containing local minima, where reactive techniques typically fail.
Coalition formation, a key factor in multi-agent cooperation, can be solved optimally for at most a few dozen agents. This paper proposes a general approach to find suboptimal solutions for a large-scale coalition formation problem containing thousands of agents using multi-agent simulation. We model coalition formation as an iterative process in which agents join and leave coalitions, and we propose several valuation functions that assign values to the coalitions. We propose several coalition selection strategies that agents may use to decide whether or not to leave their current coalition and which coalition to join. We also show how these valuation functions and coalition selection strategies represent specific coalition formation applications. Finally, we show almost-optimal performance of our algorithms in small-scale scenarios by comparing our solutions with an optimal solution, and we show stable performance in a large-scale setting in which searching for the optimal solution is not feasible.
'/%3e%3cpath d='M200 752.362h200v300H200z' style='fill:%23fff;fill-opacity:1;stroke:none' transform='translate(0 -752.362)'/%3e%3cpath d='M400 752.362h200v300H400z' style='fill:%23ff883e;fill-opacity:1;stroke:none' transform='translate(0 -752.362)'/%3e%3c/svg%3e)
