Optimization of O2O Food Delivery Strategy in Smart Cities

This paper studies an Online-to-Offline food delivery problem (OFDP) which can be viewed as a combination of variants of vehicle routing problems (VRPs). First, We define and model the OFDP mathematically. Then, we propose a novel adaptive parameters genetic algorithm with local search (APGALS) to solve the OFDP. The adaptive parameters method dynamically adjusts the crossover and mutation rates to avoid trapping into the local optimum. The local search algorithm can explore the solution space of the problem more efficiently. Static and dynamic experiments are undertaken to evaluate the performance of APGALS. The preliminary experimental results show that the adaptive parameters method and local search algorithm can improve the performance of the algorithm and the proposed APGALS is superior to the pure genetic algorithm, simulated annealing, and tabu search in terms of average fitness value and success rate in static experiment and average waiting time, number of timeout orders, and timeout accumulation in dynamic experiment.