Comparison Of Different Pick-Up And Drop-Off Modes Under The O2O Environment

Optimization of the last mile dispatching will help to reduce the cost and improve the punctuality if the last mile dispatching problem could be solved properly.There are multiple dispatching models in China O2 O market,including centralized allocation,greedy order taking by drivers and store-based delivery.Each one of these models has pros and cons.We use data provided by Cainiao's ‘The Last Mile Rush' data competition to build this Pick-up and Drop-off problem with time window.This data set contains a list of O2 O orders from merchants.We simulate different dispatching business models and make a comparison of results.We use Tabu search algorithm to generate centralized order allocation plan.We also propose greedy order-taking algorithm and a region based greedy order-taking algorithm to validate the effectiveness of Tabu search.We randomly pick ten samples from Cainiao's data set,each including 500 orders.The computational results demonstrate that centralized order allocation is better than greedy order-taking methods,and restriction on order-taking area of drivers will help improve the efficiency of greedy order-taking method.The punctuality rate of Tabu search reaches 95%,which is higher than the other two methods.The longest route of all three methods are close to each other,but the time consumed by the short route of Tabu search is much less than the other two.