The implementation of the autonomous unmanned aerial mobility is a game changer for the on-demand delivery service in the crowded urban setting. In this study, the first of its kind commercial unmanned aerial vehicle (UAV) urban delivery program in China is targeted. Different from the traditional ground pickup and delivery services, the aerial mode considers not only the time window constraints, but also the spatial conflicts incurred during the take-off and landing operations of UAVs. To obtain the optimal flying routes of the focused problem, a mixed integer programming model is formulated. Due to its inherent complexity, the optimal schedule cannot be attained within acceptable time via the off-the-shelf solvers. To help speed up the solving process, a branch-and-cut based exact algorithm is proposed, together with a series of customized valid inequalities. To further accelerate, a greedy insertion heuristic is designed to secure high-quality initial solutions. In the numerical section, it is observed that the algorithm proposed in this paper can help solve the real-life on-demand UAV delivery problem to near optimum (within 5% optimality gap) within reasonable computation time (in 5 minutes). Note to Practitioners—With the increase of labor cost, the distribution cost increases very rapidly. In the meantime, the employment of automated vehicles for logistics reshapes the landscape of the urban last-mile delivery. As an efficient courier carrier, the unmanned aerial vehicle (UAV) is trending the autonomous delivery endeavour. When integrating UAVs into the urban delivery program, practitioners need to pay special attention to the scheduling of UAVs at the operational level in addition to the hardware of the UAVs. To help solve the UAV dispatch problem, we propose an online scheduling scheme, considering the spatial conflict constraints in the actual UAV operations. And an exact algorithm is designed to accelerate the solving process. Numerical experiments demonstrate that the proposed algorithm can achieve near optimal dispatch plan with 5% optimality gap in 5 minutes. Furthermore, it is discovered that the demand pooling is an essential decision to make for UAV-based delivery. Longer pooling time can increase the UAV efficiency with more realized demand information, but too much pooling could lead to prolonged customer waiting and a low service level.

In an era of sustainable development, considerable emphasis has been put onto energy saving, environment friendly, and social welfare as well as productivity in the manufacturing sector. In this work, an unrelated parallel manufacturing setting with time-of-use (TOU) electricity price is explored, with an aim to reduce the electricity cost and increase productivity simultaneously. A nonlinear mathematical programming model is formulated to exploit the special structure of the scheduling problem, where the quadratic constraints are reformulated as second-order-cone (SOC) constraints, and several tailored cutting planes are introduced to further tighten the feasible region of the problem. Then, the original scheduling problem is transformed into several single-machine scheduling problems with TOU electricity price, which could be relaxed as a single-objective programming problem, and it could be solved rapidly via commercial solvers, such as CPLEX. Based on the optimal solution of the relaxed problem, an approximate algorithm is proposed, where a special rounding technique is employed to assign jobs to the unrelated parallel machines in a local search manner. Furthermore, a lower bound model is constructed by eliminating the nonpreemption constraint, and an iteration-based algorithm is devised to obtain the optimal solution of the lower bound problem. Meanwhile, a dispatch rule-based approach is proposed to provide an upper bound of the scheduling problem with TOU constraint. In the numerical analysis section, the proposed approximate algorithm is validated through extensive testing on various scales of instances, different emphasis on productivity and electricity price, and under two typical TOU electricity pricing policies. It is observed that the gap between the proposed approximate algorithm and CPLEX is mostly within 4%, and the lower/upper bound methods could obtain a relaxed/feasible solution within 0.01 s.