并行搜救算法在确定灾后搜救路线中的应用*

针对风暴潮辅助决策系统中对受灾区域进行灾后搜救的问题,提出了将搜救队伍扩展为两支的特殊情况下如何进行搜救路线的确定。根据总体上搜救效率达到最优的要求,在Prim算法基础上提出了并行搜救调度算法,实现了高效的搜救路线图的确定,从而使搜救行动可以高效地进行,保证了良好的辅助决策效果。     

Branch-and-price and adaptive large neighborhood search for the truck and trailer routing problem with time windows

Motivated by a situation faced by infrastructure service providers operating in urban areas with accessibility restrictions, we study the truck and trailer routing problem with time windows (TTRPTW). In this problem the vehicle fleet consists of trucks and trailers which may be decoupled. A set of customers has to be served and some of the customers can only be accessed by the truck without the trailer. This gives rise to the planning of truck-and-trailer routes containing truck-only subroutes, in addition to truck-only routes and truck-and-trailer routes without subroutes. We propose a branch-and-price algorithm for the TTRPTW, using problem specific enhancements in the pricing scheme and alternative lower bound computations. We also tailor an adaptive large neighborhood search algorithm to the TTRPTW in order to obtain good initial columns. When compared to existing metaheuristic algorithms we obtain highly competitive results. Some instances with up to 100 customers are solved to optimality with the proposed branch-and-price algorithm.     

公交最佳多路径搜索算法的应用研究

城市公交网最佳多路径搜索算法在公交出行路径查询中有着实际应用需要。针对Bus行驶路线特征提出相应的建模方法,并结合上、下行相同的公交线路模型,给出基于背离路径思想搜索次短和次次短路径的算法,提出剔除环路的不同处理策略会导致不尽相同的最佳路径。在选取上海目前66条公交线路的模型上进行了实践,证明是可行和有效的。     

电力网络机巡智能调度编排系统设计

目前,大多数国家电网公司仍采用“人巡+机巡”协调定期检查的方式对输电线路进行维护;定期检验计划由人工安排,耗费大量人力物力;为了解决上述问题,提出电力网络机巡智能调度编排系统设计,考虑到各城市年检的个性化需求,将年检计划的调度问题转化为NP问题,采用禁忌搜索算法结合遗传算法实现年检计划的智能调度;通过仿真实验验证了算法的可行性,证明了禁忌搜索算法与遗传算法的结合可以有效提高机器巡更网格的效率。     

保护区移动巡护系统

目前很多自然保护区的巡护工作仍然使用传统的巡护方式,严重影响了巡护工作效率.本文针对自然保护区的实际巡护工作需求,基于Android移动开发、Java Web开发、KML、位置服务（Location Based Service,LBS）等主要技术来构建保护区移动巡护系统,以取代传统的巡护方式.系统包括移动端采集系统和Web端管理系统.巡护人员使用移动端采集系统提高巡护信息采集的效率;保护区的管理人员使用Web系统对巡护数据进行高效管理,并及时了解巡护人员的工作情况和实时位置.系统已在多个保护区得到实际使用,并且取得了较好的应用效果.     

基于多智能体强化学习的无人机群室内辅助救援

本文主要研究了在室内场景中使用多台无人机设备对受害者进行合作搜索的问题.在室内场景中,依赖全球定位系统获取受害者位置信息可能是不可靠的.为此,本文提出一种基于多智能体强化学习(MARL)方案,该方案着重对无人机团队辅助救援时的路径规划问题进行研究.相比于传统方案,所提方案在大型室内救援场景中更具优势,例如部署多台救援无...     

路由收费清分算法

随着上海市轨道交通的不断发展,同一个轨道交通网络中将会有不同的运营商来运营其中的某些线路,由于两站点之间可能有多条路径通过,路径上可能涉及到不同运营商运营的线路,所以运营商之间的利益分配问题将会成为一个关键因素。本文给出了一种根据路由收费的清分算法,通过计算出两站点之间的所有路径以及根据这些路径上涉及到的不同运营商运营的线路,将这笔车费分配给运营商,算法中包括对站点的属性描述,起始点和终点站之间所有路径的遍历算法,各路径之间的收入分配,以及各路径上不同运营商之间的收入分配,最后给出实例以说明上述算法。     

传感器网络中基于移动sink最优穿越路径的高效数据收集算法

移动sink数据收集研究主要集中在轨迹固定的移动策略。为此,建立了移动sink穿越路径可控条件下的数据收集模型,并提出一种最优路径穿越的移动sink数据收集算法(OTP-TS)。算法利用局部Voronoi网格将连续路径问题离散化,以收集数据量、系统能耗为衡量指标,结合禁忌搜索算法,实现系统数据量最大化和网络能耗最小化穿越。理论分析和实验结果表明,该算法很好地解决了移动sink在可控移动轨迹情况中数据收集的最优路径穿越问题。     

The partial sequenced route query with traveling rules in road networks

In modern geographic information systems, route search represents an important class of queries. In route search related applications, users may want to define a number of traveling rules (traveling preferences) when they plan their trips. However, these traveling rules are not considered in most existing techniques. In this paper, we propose a novel spatial query type, the multi-rule partial sequenced route (MRPSR) query, which enables efficient trip planning with user defined traveling rules. The MRPSR query provides a unified framework that subsumes the well-known trip planning query (TPQ) and the optimal sequenced route (OSR) query. The difficulty in answering MRPSR queries lies in how to integrate multiple choices of points-of-interest (POI) with traveling rules when searching for satisfying routes. We prove that MRPSR query is NP-hard and then provide three algorithms by mapping traveling rules to an activity on vertex network. Afterwards, we extend all the proposed algorithms to road networks. By utilizing both real and synthetic POI datasets, we investigate the performance of our algorithms. The results of extensive simulations show that our algorithms are able to answer MRPSR queries effectively and efficiently with underlying road networks. Compared to the Light Optimal Route Discoverer (LORD) based brute-force solution, the response time of our algorithms is significantly reduced while the distances of the computed routes are only slightly longer than the shortest route.     

An autonomous ground robot to support firefighters' interventions in indoor emergencies

Intervention teams act in hostile scenarios where reducing mission times and accident risks is critical. In these situations, the availability of accurate information about the environment plays a key role in ensuring the well-being of rescuers and victims. This information required to plan the interventions in indoor emergencies encompasses the location of fires and the presence of dangerous gases. Robotics and remote sensing technologies can help emergency teams to obtain this information in real-time without exposing themselves. Additionally, the accurate simulation of the environments allows the teams to plan their interventions, creating routes to safely access the affected areas and evacuate the victims. This article presents a robotic solution developed to satisfy the demands of intervention teams. More specifically, it describes an autonomous ground robot that can obtain real-time location and environmental data from indoor fires, as well as a simulator that reproduces these emergency scenarios and facilitates mission planning. In this way, emergency teams can know the conditions in the scenario before, during, and after the intervention. Thus, risks are minimized by improving their situational awareness and reducing their exposure times during the mission. The system has been developed and validated in collaboration with the end-users and under realistic harsh environments. During these experiments, the robot was used to detect fire sources and cold smoke and provide environmental information to firefighters. Additionally, the simulator provided alternative routes for accessing and exiting the scene faster and safer by dodging potentially dangerous areas.     

Using auction-based task allocation scheme for simulation optimization of search and rescue in disaster relief

In order to improve the efficiency of search and rescue (SAR) in disaster relief, we optimize the SAR through agent-based simulation in this paper. The Truncated Lévy walks model is adopted to simulate rescuers’ search behaviors, and we improve it to fit the disaster environment. An auction-based task allocation scheme is used to develop a cooperative rescue plan. To verify the effectiveness of the proposed scheme, we illustrate it with the case of landslide disaster relief, and simulate it in three scenarios ("fatal", "serious" and "normal"). We compare it with non-cooperative rescue plan and the rescue plan based on well-known F-Max-Sum. The simulation results indicate that the cooperative rescue plan could improve the rescue efficiency significantly, and it performs somewhat better than the F-Max-Sum-based approach in regard to some indicators. Furthermore, its low complexity has made it more appropriate for the cooperation among rescue teams than F-Max-Sum. The robustness analysis shows that search radius can affect the rescue efficiency significantly, while the scope of cooperation has little effect on the rescue efficiency. The sensitivity analysis shows that the two parameters, the time limit for completing rescue operations in one buried site and the maximum turning angle for next step, both have great influence on rescue efficiency, and there exist optimal value for both of them in view of rescue efficiency.     

高速公路GPS车辆监控管理系统的应用

在高速公路建设取得突飞猛进的今天,针对高速公路路政巡逻、道路养护车辆的工作特点,建设一个GPS车辆监控管理系统,综合了GPS全球卫星定位系统、GIS地理信息系统、GSM／GPRS全球移动通信系统、计算机网络技术,使之能随时对车辆进行监控、调度,解决车辆管理混乱、组织无序的问题,从而对高速公路路政巡逻、抢险工作进行更加科学的管理,提高高速公路调度、指挥、管理的水平。     

多层建筑应急疏散仿真

以多层建筑为背景,研究应急疏散问题,基于改进的路径规划算法,并将多智能体技术应用在模型间的通信交流上.采用机器人感知周围环境,通过设置机器人个数以及初始位置,对灾情中室内被困人员进行智能搜救,并采集现场实时数据,作出决策分析.机器人实时感知现场状况的变化,引导人员疏散,并将实时数据传输给施救人员,采取进一步救援措施.结果表明,该三维仿真技术为有效减少人员疏散中的伤亡和最佳救援方案的制定提供了参考,具有一定的现实指导意义.     

基于JADE的海上搜救智能决策支持系统

海上搜救决策是一项复杂的系统工程,针对不同的海上突发事件,在不同时段和反应阶段,按照相关的应急预案要求,需要及时地调整决策方案.研究了基于JADE的海上搜救智能决策支持系统,描述了系统分析、系统设计、系统实现等的基本思路.结果表明,依靠海上搜救决策支持系统可以减少人为因素的干扰,对海上突发事件的危害程度、应急反应行动方...     

求解CARP-RP-ML问题的改进算法

传统方法无法有效求解交通道路维护运作中的有补给点及多装载的容量约束弧路径(CARP-RP-ML)问题。为此,提出改进的启发式算法和遗传算法。启发式算法将不同的分割算法用于由所有需求弧随机排序得到的个体上,构造问题的可行解;遗传算法利用分割算法计算其个体适应值,确定对应的可行车辆路径及补给位置,并用局部搜索作为变异算子,进一步扩大搜索空间。数值实验结果表明,与启发式算法相比,遗传算法能更有效地求解CARP-RP-ML问题。     

Hierarchical visibility for guaranteed search in large-scale outdoor terrain

Searching for moving targets in large environments is a challenging task that is relevant in several problem domains, such as capturing an invader in a camp, guarding security facilities, and searching for victims in large-scale search and rescue scenarios. The guaranteed search problem is to coordinate the search of a team of agents to guarantee the discovery of all targets. In this paper we present a self-contained solution to this problem in 2.5D real-world domains represented by digital elevation models (DEMs). We introduce hierarchical sampling on DEMs for selecting heuristically the close to minimal set of locations from which the entire surface of the DEM can be guarded. Locations are utilized to form a search graph on which search strategies for mobile agents are computed. For these strategies schedules are derived which include agent paths that are directly executable in the terrain. Presented experimental results demonstrate the performance of the method. The practical feasibility of our approach has been validated during a field experiment at the Gascola robot training site where teams of humans equipped with iPads successfully searched for adversarial and omniscient evaders. The field demonstration is the largest-scale implementation of a guaranteed search algorithm to date.     

一种结合局部搜索策略的求解TSP的演化算法

介绍了一种结合局部搜索策略的求解流动旅行商问题（TSP）的演化算法。该算法的主要思想是将局部搜索策略在邻域内搜索的快速性与演化方法在全局搜索上的鲁棒性结合起来，从而跳离局部最优。将该算法用于TSPLIB中部分TSP实例上的试验结果表明：与传统的各种求解TSP的演化方法相比，该算法在获得全局最优解的精确度上有了一定的改善。     

一种基于遗传Hopfield神经网络求解TSP问题的算法

针对Hopfield网络求解TSP问题时出现无效解和收敛性能差的问题,对约束条件能量函数进行改进,构造了一种求解TSP问题的遗传Hopfield神经网络算法,并与经典Hopfield神经网络求解TSP方法进行对比.实验结果表明,本文算法具有更好的整体求解性能.     

Evolving controllers for autonomous robot search teams

Deploying autonomous robot teams instead of humans in hazardous search and rescue missions could provide immeasurable benefits. In such operations, rescue workers often face environments where information about the physical conditions is impossible to obtain, which not only hampers the efficiency and effectiveness of the effort, but also places the rescuers in life-threatening situations. These types of risk promote the potential for using robot search teams in place of humans. This article presents the design and implementation of controllers to provide robots with appropriate behavior. The effective utilization of genetic algorithms to evolve controllers for teams of homogeneous autonomous robots for area coverage in search and rescue missions is described, along with a presentation of a robotic simulation program which was designed and developed. The main objective of this study was to contribute to efforts which attempt to implement real-world robotic solutions for search and rescue missions.     

基于进化ResNet的交通标志识别

卷积神经网络具有较优的图像特征提取性能,被广泛应用于交通标志识别领域。然而,现有交通标志识别算法通常基于专家经验设计改进的图像特征提取网络,需经历图像预处理和模型调参过程,导致模型的复杂度增大。提出一种基于进化ResNet的交通标志识别算法。将ResNet的构建参数嵌入到进化算法中,在架构搜索空间中以构建块作为基本单位,并将网络深度、卷积层通道数、池化层类型和模块构建顺序作为搜索空间的可变参数,利用交叉、变异等遗传算子执行自适应优化搜索,以确保进化搜索的有效性,同时设计适用于交通标志识别的轻量化网络。在德国交通标志数据集上的实验结果表明,该算法的识别精度达到99.41%,而参数量仅为2.37×10⁶,相比Multi-column DNN、MFC、MFC+ELM等算法,在保证识别精度的同时减少网络参数量。