Refine
Document Type
- Doctoral thesis (2)
- Bachelor thesis (1)
- Report (1)
- Working paper (1)
Has Fulltext
- yes (5)
Is part of the Bibliography
- no (5)
Keywords
- Travelling-salesman-Problem (5) (remove)
Institute
One of the standard approaches for solving discrete optimization problems which include the aspect of time, such as the traveling salesman problem with time windows, is to derive a so-called time-indexed formulation. If the problem has an underlying structure that can be described by a graph, the time-indexed formulation is usually based on a different, extended graph, commonly referred to as the time-expanded graph. The time-expanded graph can often be derived in such a way that all time constraints are incorporated in its topology, and therefore algorithms for the corresponding time-independent variant become applicable. The downside of this approach is that the sets of vertices and arcs of the time-expanded graph are much larger than the ones of the original graph. In recent works, however, it has been shown that for many practical applications a partial graph expansion that might contain time-infeasible paths, often suffices to find a proven optimal solution. These approaches, instead, iteratively refine the original graph and solve a relaxation of the time-expanded formulation in each iteration. When the solution of the current relaxation allows for a feasible schedule, an optimal solution can be derived from it and the algorithm terminates.
In this work, we first present new ideas that allow for the propagation of information about the optimal solution of a coarser graph to a more refined graph and show how these can be used in algorithms. More precisely, we present two general algorithms for solving Mixed Integer Linear Program formulations which we call iterative refinement and branch-and-refine. Iterative refinement basically is solving relaxations of the problem until a feasible solution to the original problem is found. Branch-and-refine is a kind of branch-and-bound algorithm that allows for the graph refinement to be carried out during the exploration of the branch-and-bound tree. For demonstrating the practical relevance of these algorithms, we not only study them in the context of academic examples but also apply them to two real-world problems. The first is a problem from the literature, where small passenger air-crafts have to be routed and scheduled to serve flight requests while fulfilling a variety of conditions on, for example, fuel consumption, weight, and detours. We show here that refinement algorithms can be used to improve the best known results from the literature. The second problem we consider is the task of optimally scheduling deliveries and charging times of delivery robots such that delays are minimized. In this case, we show that refinement algorithms perform better than a direct solution approach making use of state-of-the-art solvers.
Military installations and objects in out-of-area missions, e.g., an air base or a field camp, must be protected from incoming hostile rockets, artillery or mortar fire. Lasers as directed energy weapons are able to destroy those targets within seconds. Generally, the laser is assigned to a target, that applies the smallest movement of its direction unit to aim at it. The goal is to minimize the damage and thus, to destroy all incoming targets. We model the problem as a multiple traveling salesperson problem with moving targets, where the salespersons correspond to the lasers. The targets move over time on continuous trajectories. Additionally, each target is given a visibility time window. We investigate if exact methods are able to solve real-world instances in reasonable time. On that account, we address the problem from two sides, offline and online.
One essential aspect studied in this work is to find an appropriate formulation to model the time requirements. We present five different modeling approaches, where the time aspect is handled in different ways: discrete, continuous, directly or via sub-problems. Our randomly generated test instances consider 6 to 20 targets and 1 to 6 salespersons. Computational experiments with linear and non-linear trajectories are performed. The best model can solve instances up to 10 targets within 3 seconds. For online experiments the two familiar strategies REPLAN and IGNORE are adapted to our problem.
Another important aspect of this work is our contribution to competitive analysis, a method to evaluate the quality of online algorithms. Here, we restrict the problem considered so far to one salesperson and address the online moving targets traveling salesperson problem on the real line. We prove a lower bound for the competitive ratio regarding this problem. Then, we develop an online algorithm and present its competitive ratio with the corresponding proof. The competitive ratio depends on the speed ratio of salespersons and targets and outperforms a comparable online algorithm from the literature for certain speed ratios. The theoretical results obtained for the online moving target traveling salesperson problem on the real line are new in this research area.
One of the standard approaches for solving time-dependent discrete optimization problems, such as the travelling salesman problem with time-windows or the shortest path problem with time-windows is to derive a so-called time-indexed formulation. If the problem has an underlying structure that can be described by a graph, the time-indexed formulation is usually based on a different, extended graph, commonly referred to as the time-expanded graph. The time-expanded graph can often be derived in such a way that all time constraints are incorporated in its topology, and therefore algorithms for the corresponding time-independent variant become applicable. The downside of this approach is, that the sets of vertices and arcs of the time-expanded graph are much larger than the ones of the original graph. In recent works, however, it has been shown that for many practical applications a partial graph expansion, that might contain time infeasible paths, often suffices to find a proven optimal solution. These approaches, instead, iteratively refine the original graph and solve a relaxation of the time-expanded formulation in each iteration. When the solution of the current relaxation is time feasible an optimal solution can be derived from it and the algorithm terminates. In this work we present new ideas, that allow for the propagation of information about the optimal solution of a coarser graph to a more refined graph and show how these can be used in algorithms, which are based on graph refinement. More precisely we present a new algorithm for solving Mixed Integer Linear Program (MILP) formulations of time-dependent problems that allows for the graph refinement to be carried out during the exploration of the branch-and-bound tree instead of restarting whenever the optimal solution was found to be infeasible. For demonstrating the practical relevance of this algorithm we present numerical results on its application to the shortest path problem with time-windows and the traveling salesman problem with time-windows.
Eines der am intensivsten untersuchten Probleme aus dem Bereich der Optimierung stellt das sogenannte Traveling Salesman Problem (TSP, in deutsch: Handlungsreisendenproblem) dar. Die Aufgabe hinter dem Handlungsreisendenproblem besteht darin, für eine gegebene Anzahl an Orten eine Route zu entwickeln, sodass die Gesamtlänge der Route minimal ist. Zudem darf kein Ort, bis auf den ersten, mehrmals besucht werden. Ziel der vorliegenden Bachelorarbeit ist es, statistische Untersuchungen bezüglich der Tourlänge des TSP in Abhängig der Anzahl der besuchten Städte vorzunehmen, wobei der Fokus auf einer kleinen Anzahl von Städten liegt. Dazu wurde das Problem einerseits numerisch gelöst und die Verteilung der Lösungen analysiert. Für Teilaspekte wurden analytische Lösungen berechnet.
The multiple traveling salesmen problem with moving targets is a generalization of the classical traveling salesmen problem, where the targets (cities or objects) are moving over time. Additionally, for each target a visibility time window is given. The task is to find routes for several salesmen so that each target is reached exactly once within its visibility time window and the sum of all traveled distances of all salesmen is minimal. We present different modeling formulations for this TSP variant. The time requirements are modeled differently in each approach. Our goal is to examine what formulation is most suitable in terms of runtime to solve the multiple traveling salesmen problem with moving targets with exact methods. Computational experiments are carried out on randomly generated test instances to compare the different modeling approaches. The results for large-scale instances show, that the best way to model time requirements is to directly insert them into a formulation with discrete time steps.