This program conducts simulation research on the dynamic vehicle routing problem (DVRP), which is a hot research issue in the fields of operations research and combinatorial optimization. This problem is based on the static vehicle scheduling problem and is aimed at actual driving. Time is often affected by traffic accidents, weather changes and peak periods. As shown in Figure 1, find the shortest path or the vehicle route with the shortest travel time.
1. Model description
Taking the Sioux falls network (shown in Figure 2) as the simulation network
Among them, d(x,y) represents the duration from site x to site y. By setting the allocation and access sequence of customers, the total distance of the vehicle is minimized.
2. Algorithm process
Mainly used
(1) dijkstra algorithm
(2) Genetic algorithm
(3)DynaSearch algorithm
3. Algorithm simulation results
Four different types of update mechanisms are considered during the simulation process. They are: no update (mechanism 1), only update at the customer (mechanism 2), update every once in a while (mechanism 3, the time periods are set to 5 min, 10 min and 15 min respectively), and update at key points in the network (mechanism 4)．
4. Summary
This program simulates the DVRP problem, obtains the initial route arrangement by using the GA algorithm and problem, and uses the Dynasearch algorithm to adjust the route in real time. The paper conducts a numerical simulation analysis of the Sioux Falls network. The simulation results show that the new update mechanism generates The solution is better than that obtained by other update mechanisms, and the Dynasearch algorithm can quickly optimize the route in a very short time. The more unstable the transportation network is, the more advantageous it will be to adopt new strategies.