The problem of assigning a group of Unmanned Aerial Vehicles (UAVs) to perform spatially distributed tasks often requires that the tasks will be performed as quickly as possible. This problem can be defined as the Min–Max Multiple Depots Vehicle Routing Problem (MMMDVRP), which is a benchmark combinatorial optimization problem. In this problem, UAVs are assigned to service tasks so that each task is serviced once and the goal is to minimize the longest tour performed by any UAV in its motion from its initial location (depot) to the tasks and back to the depot. This problem arises in many time-critical applications, e.g. mobile targets assigned to UAVs in a military context, wildfire fighting, and disaster relief efforts in civilian applications. In this work, we formulate the problem using Mixed Integer Linear Programming (MILP) and Binary Programming and show the scalability limitation of these formulations. To improve scalability, we propose a hierarchical market-based solution (MBS). Simulation results demonstrate the ability of the MBS to solve large scale problems and obtain better costs compared with other known heuristic solution.
This work presents a methodology for real-time estimation of wildland fire growth, utilizing a fire growth model based on a set of partial differential equations for prediction, and harnessing concepts of space-time Kalman filtering and Proper Orthogonal Decomposition techniques towards low dimensional estimation of potentially large spatio-temporal states. The estimation framework is discussed in its criticality towards potential applications such as forest fire surveillance with unmanned systems equipped with onboard sensor suites. The effectiveness of the estimation process is evaluated numerically over fire growth data simulated using a well-established fire growth model described by coupled partial differential equations. The methodology is shown to be fairly accurate in estimating spatio-temporal process states through noise-ridden measurements for real-time deployability.
This work presents a methodology for real-time estimation of wildland fire growth, utilizing afire growth model based on a set of partial differential equations for prediction, and harnessing concepts of space-time Kalman filtering and Proper Orthogonal Decomposition techniques towards low dimensional estimation of potentially large spatio-temporal states. The estimation framework is discussed in its criticality towards potential applications such as forest fire surveillance with unmanned systems equipped with onboard sensor suites. The effectiveness of the estimation process is evaluated numerically over fire growth data simulated using a well-established fire growth model described by coupled partial differential equations. The methodology is shown to be fairly accurate in estimating spatio-temporal process states through noise-ridden measurements for real-time deploy ability.
Uninhabited aerial vehicles provide numerous advantages in fighting wildland fires that include persistent operation and elimination of humans from performing what can be dull, dangerous, and dirty work. Multiple cooperating uninhabited aerial vehicles can potentially bring about a paradigm shift in the way we fight complex wildland fires. This paper investigates algorithmic development for cooperative control of a number of uninhabited aerial vehicles engaged in fighting a wildland fire. The paper considers two tasks to be performed by a group of uninhabited aerial vehicles: 1) Cooperative tracking of a fire front for accurate situational awareness, and 2) cooperative, autonomous fire fighting using fire suppressant fluid. The scenario considered in this paper makes the following assumptions: information regarding the location of the fire and position of all uninhabited aerial vehicles is made available to each uninhabited aerial vehicle; and each uninhabited aerial vehicle is equipped with unlimited fire suppressant fluid which extinguishes fire in a circle of specified area directly beneath it. This paper formulates these two tasks of fire fighting based upon optimization of respective utility functions, develops a decentralized control method for the cooperative uninhabited aerial vehicles, and analyzes the system for its stability and its ability to carry out the tasks. The proposed strategies have been verified with the help of extensive simulations. Although simplifying assumptions have been made, this preliminary study presents a framework for path planning and cooperative control of multiple uninhabited aerial vehicles engaged in gathering data and actively fighting forest fires.
Tasks allocation is a fundamental problem in multiagent systems. We formulate the problem as a multiple traveling salesmen problem (MTSP), which is an extension to the well known traveling salesman problem (TSP), both considered to be NP-hard combinatorial optimization problems. We propose a solution in which agents interact in an economic market to win tasks situated in an environment. The agents strive to minimize required costs, defined as either the total distance traveled by all agents or the maximum distance traveled by any agent. Using a set of simple market operations, the agents come up with a solution for task allocation. In this work we examine the processing speed of the market-based solution (MBS), as well as the quality vs. optimal solutions achieved using enumeration for a 3 agents by 8 tasks scenario. We show that the MBS is both quick and close to optimal. We then show that the MBS can be scaled to more complicated problems, by comparing its results with results from genetic algorithm (GA) and clustering. We also show the robustness of the MBS to changes in the scenario, e.g. the addition and removal of tasks or agents.
This paper describes a market-based solution to the problem of assigning mobile agents to tasks. The problem is formulated as the multiple depots, multiple traveling salesmen problem (MTSP), where agents and tasks operate in a market to achieve near-optimal solutions. We consider both the classical MTSP, in which the sum of all tour lengths is minimized, and the Min-Max MTSP, in which the longest tour is minimized. We compare the market-based solution with direct enumeration in small scenarios, and show that the results are nearly optimal. For the classical MTSP, we compare our results to linear programming, and show that the results are within 1 % of the best cost found by linear programming in more than 90 % of the runs, with a significant reduction in runtime. For the Min-Max case, we compare our method with Carlsson's algorithm and show an improvement of 5 % to 40 % in cost, albeit at an increase in runtime. Finally, we demonstrate the ability of the market-based solution to deal with changes in the scenario, e.g., agents leaving and entering the market. We show that the market paradigm is ideal for dealing with these changes during runtime, without the need to restart the algorithm, and that the solution reacts to the new scenarios in a quick and near-optimal way.
Wildfire is one of the most significant disturbances responsible for reshaping the terrain and changing the ecosystem of a particular region. Its detrimental effects on environment as well as human lives and properties, and growing trend in terms of frequency and intensity of wildfires over the last decade have necessitated the development of efficient forest fire management techniques. During the last three decades, Forest Fire Decision Support Systems (FFDSS) have been developed to help in the decision-making processes during forest fires by providing necessary information on fire detection, their status and behavior, and other aspects of forest fires. However, most of these decision support systems lack the capability of developing intelligent fire suppression strategies based upon current status and predicted behavior of forest fire. This paper presents an approach for development of efficient fireline building strategies via intelligent resource allocation. A Genetic Algorithm based approach has been proposed in this paper for resource allocation and optimum fireline building that minimizes the total damage due to wildland fires. The approach is based on a simulation–optimization technique in which the Genetic Algorithm uses advanced forest fire propagation models based upon Huygens principles for evaluation of cost index of its solutions. Both homogeneous and heterogeneous environmental conditions have been considered. Uncertainties in weather conditions as well as imperfect knowledge about exact vegetation and topographical conditions make exact prediction of wildfires very difficult. The paper incorporates Monte-Carlo simulations to develop robust strategies in uncertain conditions. Extensive simulations demonstrate the effectiveness of the proposed approach in efficient resource allocation for fighting
complex wildfires in uncertain and dynamic conditions.