On The Optimization Method And System For The Dynamic Scheduling Of Transport Vehicles

As the strategy of strengthening the army with science and technology isincreasingly intensified, the dynamic optimization scheduling for transport vehicles hasbecome a hot issue of military transport in the field of Chinese military logistics.However, modern warfare has proved a reality that the objects of military transportinclude more than the logistical materials mentioned in military logistics, also otherelements such as soldiers, ammunition and equipment, which contribute to the combateffectiveness. These special objects require the types of vehicles responsible for thetransport of these elements to be various, instead of just material transporter. In addition,ever-changing battlefield environment calls for rapid dynamic scheduling of theseelements, thus increases the frequency of transport vehicles’ dynamic scheduling. Largescale of tasks, various types of vehicles and high dynamics have been a difficult problemto the department of transport vehicle scheduling which undertakes the task of Chinesemilitary transport. How to realize the dynamic optimization scheduling of large-scale andmultiple transport vehicles is crucial to strengthening Chinese army’s ability to win inconditions of modern high-tech war. This thesis, set in large-scale military exercises,probes into the dynamic optimization scheduling of transport vehicles in land militarytransport.Starting with dynamic scheduling of transport vehicles from home and abroad, thisthesis analyzes and summarizes the achievements in this field, as well as further researchdirections. Furthermore, it briefly introduced the systems for the dynamic scheduling oftransport vehicles applied into operation at home and abroad.This thesis makes a comparatively complete elaboration on the basic approachandsloving idea of Approximate Dynamic Programming (ADP) as a whole, then severalkey concepts of Decision Function, Sample Path and Approximate Value Function,involved in the application of ADP for problem solving in this thesis, are furtherinterpreted in theory. With descriptive parameters of ADP as the starting point, thisthesis introduces a common method in the assessment of the ADP-aggregation, on the basis of which two common algorithm frameworks of ADP are premised. Finally, thethesis introduces the application scope of ADP.Then beginning with the features of large-scale and multiple vehicle dynamicscheduling, this thesis gives its definition and models vehicle resources, transportationtasks, scheduling decisions, exogenous information and sampling time in it on the basisof the thought of ADP. In the view of the features of such problems, the heuristic (myopic)policy as solution to it is put forward. Combining the model with heuristic policy, thesolution algorithm is presented based on ADP, to which the thesis makes a simulationexperiment and data analysis to verify the correctness of the algorithm. The greedyalgorithms based on the heuristic policy, superiority of algorithm is verified.The management system for the dynamic scheduling of transport vehicles getsdesigned and developed in the framework of CERP3X. Furthermore, the algorithm forthe dynamic scheduling of transport vehicles based on ADP is applied to the system so asto complete the implementation and verification of the algorithm in it. And then thepotential application of the system is discussed and the procedure of application isillustrated.