Abstract: According to whether customers purchase on-time delivery services or are willing to pay additional fees for early delivery, orders with instant delivery are classified into different types. In addition to the initial order requirements, new order requirements may also arise during the delivery process. A mathematical model was established with the goal of minimizing the sum of vehicle delivery costs and time costs at customer points, taking into account constraints such as the time window of orders, vehicle capacity limitations, and crowdsourcing vehicle service scope. An improved hybrid tabu search algorithm based on rolling time domain was designed to solve the problem, which includes a dynamic adjustment mechanism for tabu step size and a diversification strategy for solutions. Parameter analysis shows that in order to effectively reduce costs, transportation companies should not set the update interval too long. Instead, they should prioritize the delivery of heterogeneous orders of the second and third categories, expand the service scope of crowdsourced vehicles as much as possible, and fully utilize the crowdsourced vehicles within this range. Multiple case studies of different scales have shown that the improved hybrid tabu search algorithm based on rolling time domain can effectively solve cases of various scales.