Research On Intelligent Scheduling Optimization And Task Cooperation In Complex Environment

With the rapid development of 5G and Internet of things technology,agent devices with environmental awareness gradually play an irreplaceable role.In a variety of production and life scenarios,with the help of the networking interaction and self decision-making ability of this kind of equipment,we can obtain higher production efficiency under many complex constraint scenarios that can not be dealt with before,which greatly reduces the production and living costs,and has high research value.In this paper,we consider the application scenarios with complex terrain,and study two different intelligent scheduling problems.Firstly,this paper considers the intelligent bus scheduling problem in complex road network.In modern cities,the passenger volume of urban subway often exceeds the warning threshold,and the passenger satisfaction tends to decline.How to improve the punctuality rate of buses so that more passengers can use buses instead of private cars is an interesting and important research.With the continuous emergence of connected automatic vehicles(CAVs)and vehicle to infrastructure(V2I)in intelligent transportation system,improving the punctuality rate of public transportation has become the goal of modern public transportation system.Considering the time-varying passenger flow and real-time road congestion,this paper optimizes the signal timing scheme under multiple bus routes and the bus speed scheme on the road to minimize the total waiting time of passengers in the bus system.In order to solve this problem,we propose a new scheme which combines TSSA with real-time scheduling framework.TSSA has the following characteristics:Ⅰ)with the increase of TSSA production,a variety of combinations of chain sheath,linear sheath,cluster sheath and spiral sheath were produced through grouping activities.This is the cornerstone of improving search performance.Ⅱ)the circular and spiral curve equations are used to enhance the search.These equations are used in the cluster and spiral bottle models to avoid premature convergence.Based on the real bus data,20 test questions(such as 6 bus lines,115 intersections,321 stops and 197 roads)are constructed.Compared with the four comparison algorithms,the effectiveness and robustness of TSSA in improving bus punctuality are verified.Secondly,the distributed multi robot cooperative optimization simulation under limited communication is considered.The foraging of multi robot system in unknown environment is a challenging problem in recent years.Robot team must rely on effective coordination strategies to deal with the uncertainty and danger brought by unknown environment information.With the help of communication between members,collision and task loss caused by environmental changes and single point of failure can be effectively reduced,but a large amount of information exchange is also accompanied The increase of time and energy cost reduces the efficiency of foraging.In order to solve this problem,this paper studies the cooperative foraging problem of multi robot system with limited communication range in unknown environment.Our mathematical model is based on grid network and distributed control architecture,and considers the change of communication topology caused by limited communication range and the task allocation relationship in local network.In order to reduce the cost of limited information sharing,this paper designs a distributed cooperative foraging algorithm VFAA based on virtual force and parallel auction.The concept of virtual force is used to guide the movement of robot members,so as to maintain local communication links in a limited step size and reduce the uncertainty of information.The local task resources are auctioned and replanning in the local sub network to obtain the optimal solution under the local path and reduce the local driving cost of individual members.In order to verify the effectiveness of the proposed algorithm,we carried out simulation experiments in 14 test sets of different scales.The results show that,compared with the other four typical robot foraging methods,the proposed algorithm obtains more foraging amount and lower execution cost in most test sets,showing good performance.