Vehicle Routing Problems that Minimize the Completion Time: Heuristics, Worst-Case Analyses, and Computational Results

In the standard Vehicle Routing Problem (VRP), we route a fleet of vehicles to deliver the demands of all customers such that the total distance traveled by the fleet is minimized. In this dissertation, we study variants of the VRP that minimize the completion time, i.e., we minimize the distance of the longest route. We call it the min-max objective function. In applications such as disaster relief efforts and military operations, the objective is often to finish the delivery or the task as soon as possible, not to plan routes with the minimum total distance. Even in commercial package delivery nowadays, companies are investing in new technologies to speed up delivery instead of focusing merely on the min-sum objective. In this dissertation, we compare the min-max and the standard (min-sum) objective functions in a worst-case analysis to show that the optimal solution with respect to one objective function can be very poor with respect to the other. The results motivate the design of algorithms specifically for the min-max objective. We study variants of min-max VRPs including one problem from the literature (the min-max Multi-Depot VRP) and two new problems (the min-max Split Delivery Multi-Depot VRP with Minimum Service Requirement and the min-max Close-Enough VRP). We develop heuristics to solve these three problems. We compare the results produced by our heuristics to the best-known solutions in the literature and find that our algorithms are effective. In the case where benchmark instances are not available, we generate instances whose near-optimal solutions can be estimated based on geometry. We formulate the Vehicle Routing Problem with Drones and carry out a theoretical analysis to show the maximum benefit from using drones in addition to trucks to reduce delivery time. The speed-up ratio depends on the number of drones loaded onto one truck and the speed of the drone relative to the speed of the truck.

Exploiting Nested Parallelism on Heterogeneous Processors

Heterogeneous computing systems have become common in modern processor architectures. These systems, such as those released by AMD, Intel, and Nvidia, include both CPU and GPU cores on a single die available with reduced communication overhead compared to their discrete predecessors. Currently, discrete CPU/GPU systems are limited, requiring larger, regular, highly-parallel workloads to overcome the communication costs of the system. Without the traditional communication delay assumed between GPUs and CPUs, we believe non-traditional workloads could be targeted for GPU execution. Specifically, this thesis focuses on the execution model of nested parallel workloads on heterogeneous systems. We have designed a simulation flow which utilizes widely used CPU and GPU simulators to model heterogeneous computing architectures. We then applied this simulator to non-traditional GPU workloads using different execution models. We also have proposed a new execution model for nested parallelism allowing users to exploit these heterogeneous systems to reduce execution time.