Kreß, Dominik

Constraint programming solvers are known to perform remarkably well for most scheduling problems. However, when comparing the performance of different available solvers, there is usually no clear winner over all relevant problem instances. This gives rise to the question of how to select a promising solver when knowing the concrete instance to be solved. In this article, we aim to provide first insights into this question for the flexible job shop scheduling problem. We investigate relative performance differences among five constraint programming solvers on problem instances taken from the literature as well as randomly generated problem instances. These solvers include commercial and non-commercial software and represent the state-of-the-art as identified in the relevant literature. We find that two solvers, the IBM ILOG CPLEX CP Optimizer and Google’s OR-Tools, outperform alternative solvers. These two solvers show complementary strengths regarding their ability to determine provably optimal solutions within practically reasonable time limits and their ability to quickly determine high quality feasible solutions across different test instances. Hence, we leverage the resulting performance complementarity by proposing algorithm selection approaches that predict the best solver for a given problem instance based on instance features or parameters. The approaches are based on two machine learning techniques, decision trees and deep neural networks, in various variants. In a computational study, we analyze the performance of the resulting algorithm selection models and show that our approaches outperform the use of a single solver and should thus be considered as a relevant tool by decision makers in practice.

We consider a semiconductor final-test scheduling problem that aims at minimizing the total weighted tardiness. In contrast to previous studies on this problem, we explicitly take account of the need to assign human operators to setup operations. We present decomposition-based heuristic solution approaches and a mixed integer program. In a computational study based on real-world problem instances that mimic settings at our industry partner, we show that our heuristics clearly outperform a standard solver when computational time is limited. Based on this result, we provide decision support for managers by analyzing the capability and effect of rescheduling jobs in the presence of a highly dynamic environment with frequently changing customer requests and common test machine failures.

This paper addresses a flexible job shop scheduling problem that takes account of workforce constraints and aims to minimise the makespan. The former constraints ensure that eligible workers that operate the machines and may be heterogeneously qualified, are assigned to the machines during the processing of operations. We develop different variants of filter-and-fan (F&F) based heuristic solution approaches that combine a local search procedure with a tree search procedure. The former procedure is used to obtain local optima, while the latter procedure generates compound transitions in order to explore larger neighbourhoods. In order to be able to adapt neighbourhood structures that have formerly shown to perform well when workforce restrictions are not considered, we decompose the problem into two components for decisions on machine allocation and sequencing and decisions on worker assignment, respectively. Based on this idea, we develop multiple definitions of neighbourhoods that are successively locked and unlocked during runtime of the F&F heuristics. In a computational study, we show that our solution approaches are competitive when compared with the use of a standard constraint programming solver and that they outperform state-of-the-art heuristic approaches on average.

This paper treats the Piggyback Transportation Problem: A large vehicle moves successive batches of small vehicles from a depot to a single launching point. Here, the small vehicles depart toward assigned customers, supply shipments, and return to the depot. Once the large vehicle has returned and another batch of small vehicles has been loaded at the depot, the process repeats until all customers are serviced. With autonomous driving on the verge of practical application, this general setting occurs whenever small autonomous delivery vehicles with limited operating range, e.g., unmanned aerial vehicles (drones) or delivery robots, need to be brought in the proximity of the customers by a larger vehicle, e.g., a truck. We aim at the most elementary decision problem in this context, which is inspired by Amazon’s novel last-mile concept, the flying warehouse. According to this concept, drones are launched from a flying warehouse and – after their return to an earthbound depot – are resupplied to the flying warehouse by an air shuttle. We formulate the Piggyback Transportation Problem, investigate its computational complexity, and derive suited solution procedures. From a theoretical perspective, we prove different important structural problem properties. From a practical point of view, we explore the impact of the two main cost drivers, the capacity of the large vehicle and the fleet size of small vehicles, on service quality.

There is a finite number of non-cooperating clients, who are averse to uncertain loss and compete for execution of their jobs not later than by their respective due dates in a parallel service environment. For each client, a due date violation implies a cost. In order to address the minimization of the total scheduling cost of all clients as a social criterion, a game mechanism is suggested. It is designed such that no client has an incentive to claim a false due date or cost. The game mechanism allows the clients to move their jobs to complete earlier in a given schedule. However, they must compensate costs of those clients whose jobs miss their due dates because of these moves. Algorithmic aspects are analyzed. Furthermore, a polynomial time algorithm that determines an equilibrium of the considered game is suggested and embedded into the game mechanism. Computational tests analyze the performance and practical suitability of the resulting game mechanism.

We consider a scheduling problem for two gantry cranes moving on the same rails at a single storage block. Containers originating at the seaside have to be stored in the block and containers that are already stored in the storage area at the beginning of the planning horizon have to be delivered to the landside handover point within given time windows. Most commonly in seaport operations, the berthing time of vessels is to be minimized. Thus, the objective considered in this article is to minimize the makespan of seaside container processing while guaranteeing on-time processing of landside containers and while considering non-crossing constraints among cranes. We allow preemption of seaside container processing. This means that one crane may move a seaside container to an intermediate storage slot, and the other crane takes it to its designated position. This has previously been shown to be an effective method of reducing the makespan when compared to classical approaches. We present a dynamic programming (DP) algorithm and a related beam search heuristic. The DP method makes use of bounding techniques and applies dominance properties of optimal solutions. In computational tests, we show that the DP approach clearly outperforms CPLEX and that it is able to quickly solve instances with real-world yard settings. The beam search heuristic is shown to be capable of quickly improving solutions of heuristic approaches that have previously been introduced in the literature. This allows both algorithms to be applied in real-world online settings, where container data is revealed incrementally.

We address an optimization problem that arises at seaports where containers are transported between stacking areas and small buffer areas of restricted capacity that are located within the reach of quay cranes. The containers are transported by straddle carriers that have to be routed such that given unloading and loading sequences of the containers at the quay cranes are respected. The objective is to minimize the turnaround times of the vessels. We analyze the problem’s computational complexity, present an integer program, and propose a heuristic framework that is based on decomposing the problem into its routing component and a component that handles the time variables and buffer capacities. The framework is analyzed in computational tests that are based on real-world data. Based on these tests, we analyze the question of whether or not it pays off to deviate from the approach of permanently assigning a fixed number of straddle carriers to each quay crane, which is the strategy that is currently implemented at the port.