Minibus Route Planning Based On A Class Of Intelligent Synthesis Algorithms

Due to the maturity of high-speed rail technology,the long-distance transportation of inter-provincial passengers has been changed from the original long-distance bus to fixed-point and scheduled delivery of high-speed rail,which not only saves costs,but also greatly saves travel time for passengers.However,if it adopts the high-speed rail method,the transportation of short-distance passengers in the province seems to be an overkill,and at the same time,it cannot send passengers to their corresponding precise locations.Therefore,this thesis proposes an efficient,accurate and flexible way of minibuses transporting in urban areas in order to meet the needs of shortdistance transportation of passengers in the province,that is,regular and fixedtime minibus passenger transport in urban areas.Therefore,we propose an improved Ant colony and Simulated annealing optimization algorithm.Based on the existing classic Ant colony algorithm solving the traditional traveling salesman problem,the improved ant colony algorithm abstracts the minibus route planning problem to be a traveling salesman-like problem with fixed starting and ending points.In order to solve the shortest path problem of this kind of urban transfer,we separate the starting position and the ending position separately,not as the iterative object of the entire algorithm,based on the original ant colony optimization algorithm.In this way,when the ants are randomly placed at any distribution point,if it is found that the next iteration of the placed ants includes the starting point and end point that we have set in advance,the iteration is forcibly abandoned.Then we reselect the corresponding iterative position for this batch of ants until we find a position that meets the requirements(the iterative position of the randomly placed ant points does not include the start and end points).Essentially,it can be summarized that it only iterates the path sequence passing points,and finally adds stored starting and ending points to it when calculating the objective function.The simulated annealing algorithm is similar to ACO.The starting and ending iteration positions are automatically excluded when the initial path and the current temperature are randomly generated.The explanation is that the iteration is cancelled and the selection is made again until the conditions are met when the path generated by the iteration includes any position of the starting point or the end point on the algorithm program.Since the improved ant colony algorithm still cannot get rid of the disadvantages of the traditional ant colony optimization algorithm.The disadvantages are that it is easy to fall into a local dead zone and the search speed is slow.So a simulated annealing algorithm is introduced to solve this problem.The simulated annealing algorithm is simpler than the ant colony optimization algorithm in the trial process.It is more inclined to optimize the local selection first,and then to screen larger localities in the form of probability.This method is relatively simple and easy to understand compared with the previous ant colony algorithm,but the simulated annealing algorithm has a narrower search range and is not capable of searching the global optimal solution.In the end,these two algorithms can solve the problem of route planning in terms of reasonably transferring passengers because of their unique advantages.However,due to their advantages and disadvantages we decided to use them together to make up for their respective disadvantages after comparing the two algorithms.We developed an ant colony-simulated annealing algorithm to comprehensively solve the shortest path planning problem of minibuses in order to make up for the ant colony algorithm's shortcomings of being easy to get in the dead zone,and insufficient accuracy,and simulated annealing algorithm's shortcomings of small searching range and weak searching power.The new algorithm draws on the deep searching capability of the simulated annealing algorithm and the bread searching capability of the ant colony algorithm.It is more comprehensive and broader than the simulated annealing algorithm.At the same time,this algorithm finds the global optimal solution with a higher probability than the ant colony algorithm,and enhances the local accuracy,that is,it enhances the depth search ability.