Study On Dynamic Vehicle Scheduling Problem With Full-Load

Vehicle scheduling problem (VSP) is the core content of cargo transportation and logisticsdistribution for transportation and logistics enterprises.The study can reduce the cost oftransportation and logistics systems by using the results of VSP. The routes of vehicles may beadopted in concrete driving process for the uncertainty of influence factors i.e. the change ofcustomer demand, weather, road jamming and congestion.With the rapid development ofinformation technology and the Internet of Things, technical support and the possibility of theimplementation are provided for dynamic vehicle scheduling problem, the study with importanttheoretical innovation and engineering application value are based on the real-time information.Based on the work of consultation and analysis at home and abroad, the vehicle schedulingproblem with full-load (VSPFL) that based on the real-time information is described in this paper.The environmental influence factors, which include road network conditions, vehicles anddrivers status, customer demand information are analyzed systematically. And the variableswithin the system, the relationship with them and system goals are discussed. Dynamic vehiclescheduling system with full-load is designed, which includes information sub-system, databasesub-system, off-line scheduling sub-system and on-line scheduling sub-system. Thesesub-systems are connected with the real-time information of road and customer. The functionsand relationship of these sub-systems and the dynamic network vehicle scheduling system aredescribed briefly.The preliminary vehicle scheduling program (scheduling plan) is determined in off-linesystem and the ultimate vehicle scheduling program with real-time information is adjusted inon-line system. Then the relevant mathematical models of these sub-systems are established, andthe respective algorithms to solve the problem are provided: Based on the historical data andadvanced customer demand, the model of off-line sub-system for solution is optimized before theactual scheduling process by using the improved ant colony algorithm (ACO), then thepreliminary scheduling program is determined (vehicle scheduling plan with full-load); Based onthe real-time information of customer demand and road, the model of on-line sub-system forsolution is optimized during the actual scheduling process by combining with the concept ofnetwork route, then the vehicle driving routes are adjusted.According to the change of road resistance and the simultaneous change of customerdemand and road resistance, VSPFL in the inter-city (or urban-rural) road is discussed and the stimulation experiments are carried out. The feasibility and validity of the theories and methodsare demonstrated by the stimulation experiment, which shows that the shortest routes arechoosed, the vehicles on the road is fully used to satisfy the dynamic change of customer demand,and the numbers of vehicle and the total transportation costs can be reduced.