In this paper we consider the integration of the personnel scheduling into planning railway yards. This involves an extension of the Train Unit Shunting Problem, in which a conflict-free schedule of all activities at the yard has to be constructed. As the yards often consist of several kilometers of railway track, the main challenge in finding efficient staff schedules arises from the potentially large walking distances between activities. We present two efficient heuristics for staff assignment. These methods are integrated into a local search framework to find feasible solutions to the Train Unit Shunting Problem with staff requirements. To the best of our knowledge, this is the first algorithm to solve the complete version of this problem. Additionally, we propose a dynamic programming method to assign staff members as passengers to train movements to reduce their walking time. Furthermore, we describe several ILP-based approaches to find a feasible solution of the staff assignment problem with maximum robustness, which solution we use to evaluate the quality of the solutions produced by the heuristics. On a set of 300 instances of the train unit shunting problem with staff scheduling on a real-world railway yard, the best-performing heuristic integrated into the local search approach solves 97% of the instances within three minutes on average.

The general problem of scheduling activities subject to temporal and resource constraints as well as a deadline emerges naturally in numerous application domains such as project management, production planning, and public transport. The schedules often have to be implemented in an uncertain environment, where disturbances cause deviations in the duration, release date or deadline of activities. Since these disruptions are not known in the planning phase, we must have schedules that are robust, i.e., capable of absorbing the disturbances without large deteriorations of the solution quality. Due to the complexity of computing the robustness of a schedule directly, many surrogate robustness measures have been proposed in literature. In this paper, we propose new robustness measures, and compare these and several existing measures with the results of a simulation study to determine which measures can be applied in practice to obtain good approximations of the true robustness of a schedule with deadlines. The experiments are performed on schedules generated for real-world scheduling problems at the shunting yards of the Dutch Railways (NS).

A multi-layered environment (MLE) [van Toll et al. 2011] is a representation of the walkable environment (WE) in a 3D virtual environment that comprises a set of two-dimensional layers together with the locations where the different layers touch, which are called connections. This representation can be used for crowd simulations, e.g. to determine evacuation times in complex buildings, or for finding the shortest routes. The running times of these algorithms depend on the number of connections.

The multiprocessor flow-shop is the generalization of the flow-shop in which each machine is replaced by a set of identical machines. As finding a minimum-length schedule is NP-hard, we set out to find good lower and upper bounds. The lower bounds are based on relaxation of the capacities of all machine sets except one. This results in a parallel-machine scheduling problem with release dates and delivery times, for which we derive a number of lower bounds. We pay special attention to the time complexity of algorithms for computing these bounds. To obtain the upper bounds a constructive algorithm in subsequent stages is used. We present an experimental comparison of the various lower and upper bounds for the multiprocessor flow-shop problem.

Column generation has proved to be an effective technique for solving the linear programming relaxation ofhuge set covering or set partitioning problems, and column generation approaches have led to state-of-the-art so-called branch-and-price algorithms for various archetypical combinatorial optimization problems. We use a combination of column generation and Lagrangean relaxation to tackle a single-machine common due date problem, where Lagrangean relaxation is exploited for early termination of the column generation algorithm and for speeding up the pricing algorithm. We show that the Lagrangean lower bound dominates the lower bound that can be derived from the column generation algorithm when applied to the standard linear programming formulation, but we also show how the linear programming formulation can be adapted such that the corresponding lower bound is equal to the Lagrangean lower bound. Our comprehensive computational study shows that the combined algorithm is by far superior to two existing purely column generation algorithms: it solves instances with up to 125 jobs to optimality, while a purely column generation algorithm can solve instances with up to only 60 jobs.

We provide several non-approximability results for deterministic scheduling problems whose objective is to minimize the total job completion time. Unless 𝒫 =𝒩𝒫, none of the problems under consideration can be approximated in polynomial time within arbitrarily good precision. Most of our results are derived by APX-hardness proofs. We show that, whereas scheduling on unrelated machines with unit weights is polynomially solvable, the problem becomes APX-hard if release dates or weights are added. We further show APX-hardness for scheduling in flow shops, job shops, and open shops. We also investigate the problems of scheduling on parallel machines with precedence constraints and unit processing times, and two variants of the latter problem with unit communication delays; for these problems we provide lower bounds on the worst-case behavior of any polynomial-time approximation algorithm through the gap-reduction technique.

We consider a single-machine on-line scheduling problem where jobs arrive over time. A set of independent jobs has to be scheduled on the machine, where preemption is not allowed and the number of jobs is unknown in advance. Each job becomes available at its release date, which is not known in advance, and its characteristics, i.e., processing requirement and delivery time, become known at its arrival. The objective is to minimize the time by which all jobs have been delivered. We propose and analyze an on-line algorithm based on the following idea: As soon as the machine becomes available for processing, choose an available job with highest priority, and schedule it if its processing requirement is not too large. Otherwise, postpone the start of this job. We prove that our algorithm has performance bound $(\sqrt{5}+1)/2 \approx 1.61803$, and we show that there cannot exist a deterministic on-line algorithm with a better performance ratio for this problem.

The just-in-time concept in manufacturing has aroused interest in machine scheduling problems with earliness-tardiness penalties. In particular, common due date problems, which are structurally less complicated than problems with general due dates, have emerged as an interesting and fruitful field of research. We prove that so-called almost common due date problems, in which the due date dj and processing time pj of each job Jj(j = 1, …, n) are such that dj ∈ [D, D + pj] for some constant D, are structurally less complicated also. Our main contribution is an O(n2) time dynamic programming algorithm for the almost common due date problem with large D. The dynamic programming algorithm is interesting in its own right, since the optimality principle behind it applies to other common due date and almost common due date problems as well.

We address the NP-hard single-machine problem of scheduling n independent jobs so as to minimize the sum of α times total completion time and β times total earliness with β > α, which can be rewritten as an earliness–tardiness problem. Postponing jobs by leaving the machine idle may then be advantageous. The allowance of machine idle time between the execution of jobs singles out our problem from most concurrent research on problems with earliness penalties. Solving the problem to optimality poses a computational challenge, since the possibility of leaving the machine idle has a major effect on designing a branch-and-bound algorithm in general, and on computing lower bounds in particular. We present a branch-and-bound algorithm which is based upon many dominance rules and various lower bound approaches, including relaxation of the machine capacity, data manipulation, and Lagrangian relaxation. The algorithm is shown to solve small instances with up to 20 jobs.