Research On Optimization Of Drones Scheduling For Ships Exhaust Emission Detection

As an important logistics mode in international trade,maritime transport drives the development of shipping industry and the operation and management of ports.The problem of ship exhaust pollution accelerates global warming and threatens human life and health in coastal areas,which constantly triggers the global port cities to strengthen their supervision.The International Maritime Organization(IMO)and various regional government departments develop and implement the relevant laws and regulations on the emission reduction of ships,among which the emission control area(ECA)is widely promoted and applied as a powerful monitoring means.The ECA policy is aimed at regulating the use of low-sulphur fuel oil standards for ships operating or berthing within the waters of the port.The use of low-sulfur fuel oil by ships will generate additional navigation costs,and the relaxation of the current monitoring means will easily trigger many shipowners to violate the rules.The remote sensing detection system,drones are equipped with exhaust sniffing technology or optical sensing technology can realize the close monitoring of multiple moving ships,which will improve the efficiency of ship exhaust emission detection and have practical application value.Based on the above background,this study takes ECA ship exhaust emission detection as the research object and describes it as drones scheduling optimization problem.In the study of drone scheduling optimization,this paper focuses on the relative quantitative relationship between ECA ship collection to be tested and drone resources,and carries out the following research from the perspective of optimization of drone resources at port aircraft stations.Based on drone resource saving,drone cooperative allocation under multi-station strategic cooperation is studied.Due to the limitation of drone resources,the problem of singlestation drone scheduling based on route optimization is studied in tactical layer.Combined with the optimization of drone resources at strategic level and tactical level,the cooperative scheduling problem of multi-station drone is studied.For ECA ships to be tested in the real scene,this paper extends the assumed static state to dynamic moving state under continuous time,and proposes a synchronous encounter model between drones and ships in the operational layer.The construction of synchronous encounter model is based on the automatic identification system(AIS),which introduces the speed of the ship,describes its navigation direction and navigation trajectory,and determines the optimal path for the drone to synchronize time and space with the real-time position of the ship to be measured.In order to verify the effectiveness of the synchronous encounter model in route optimization,the two-stage pursuit path and synchronous encounter path are described and compared to analyze the difference of flight time cost between the two paths and their influence on the optimal decision of multi-aircraft station cooperative allocation.In order to illustrate the high efficiency and applicability of synchronous encounter model in drone detection of multi-ship cyclic path optimization,drones scheduling optimization problem based on synchronous encounter cyclic path is proposed.Based on the limitation degree of the number of drone resources and the cooperation relationship between multiple aircraft and stations in the real scene,different cyclic path forms based on typical combinatorial optimization problems are designed to build complex scheduling scenarios.The drone scheduling optimization problem is subdivided into the drone scheduling optimization based on synchronous encounter TSP path,synchronous encounter One-Depot VRP path,synchronous encounter Multi-Depot TSP path and synchronous encounter Multi-Depot VRP path.In the construction of drone scheduling model based on synchronous encounter cyclic path,the drone scheduling model of cyclic path is taken as the benchmark model,the synchronous encounter model is taken as the mechanism model,integrated into the benchmark model and its extended model is established.During the flight of the drone to perform the ship exhaust emission detection,all ECA ships under test move synchronously in time and space according to their speed,size and direction,causing the spatial distance matrix between different ships to change with time.At the same time,considering the heterogeneity of the flight speed of drones at different stations,the flight time of drones at different stations performing the same ship task is different.The time attribute and interaction of the above two factors lead to the nonlinearity and complexity of the drone flight duration minimization scheduling model,which is not conducive to the use of accurate algorithms.In order to effectively solve the drone scheduling optimization problem based on synchronous encounter cyclic path,three algorithms are designed in this paper,which are sequence insertion algorithm(SIA)based on cyclic path,two-stage algorithm combining basic model solving and sequence insertion decoding(BSA),and Genetic algorithm based on sequence insertion decoding(SIA-GA).For SIA-GA with global parallel search,in order to improve its solving performance,a variety of decomposition strategies considering the characteristics of the problem are designed.The results show that the three algorithms can obtain the scheduling decision scheme based on synchronous encounter cyclic path by constructing the path sequence of a single drone visiting a ship with minimum granularity,and the SIA-GA based on population optimization has the best performance in the quality of largescale problem solving.