This dissertation discusses two independent topics: a resource-constrained shortest-path problem (RCSP) and a literature review on scheduling problems involving sequence-dependent setup (SDS) times (costs). RCSP is often used as a subproblem in column generation because it can be used to solve many practical problems. This dissertation studies RCSP with multiple resource constraints on an acyclic graph, because many applications involve this configuration, especially in column genetation formulations. In particular, this research focuses on a dynamic RCSP since, as a subproblem in column generation, objective function coefficients are updated using new values of dual variables at each iteration. This dissertation proposes a pseudo-polynomial solution method for solving the dynamic RCSP by exploiting the special structure of an acyclic graph with the goal of effectively reoptimizing RCSP in the context of column generation. This method uses a one-time “preliminary” phase to transform RCSP into an unconstrained shortest path problem (SPP) and then solves the resulting SPP after new values of dual variables are used to update objective function coefficients (i.e., reduced costs) at each iteration. Network reduction techniques are considered to remove some nodes and/or arcs permanently in the preliminary phase. Specified techniques are explored to reoptimize when only several coefficients change and for dealing with forbidden and prescribed arcs in the context of a column generation/branch-and-bound approach. As a benchmark method, a label-setting algorithm is also proposed. Computational tests are designed to show the effectiveness of the proposed algorithms and procedures. This dissertation also gives a literature review related to the class of scheduling problems that involve SDS times (costs), an important consideration in many practical applications. It focuses on papers published within the last decade, addressing a variety of machine configurations - single machine, parallel machine, flow shop, and job shop - reviewing both optimizing and heuristic solution methods in each category. Since lot-sizing is so intimately related to scheduling, this dissertation reviews work that integrates these issues in relationship to each configuration. This dissertation provides a perspective of this line of research, gives conclusions, and discusses fertile research opportunities posed by this class of scheduling problems. since, as a subproblem in column generation, objective function coefficients are updated using new values of dual variables at each iteration. This dissertation proposes a pseudo-polynomial solution method for solving the dynamic RCSP by exploiting the special structure of an acyclic graph with the goal of effectively reoptimizing RCSP in the context of column generation. This method uses a one-time

TRANSIMS (Transportation Analysis Simulation System) is part of a multi-track Travel Model Improvement Program sponsored by the U. S. Department of Transportation (DOT), and the Environmental Protection Agency (EPA). The main objective of this thesis is to enhance and implement a principal module in TRANSIMS, called the Route Planner Module. The purpose of the Route Planner Module is to find time-dependent label-constrained shortest paths for transportation activities performed by travelers in the system. There are several variations of shortest path problems and algorithms that vary by application, contexts, complexity, required data, and computer implementation techniques. In general, these variants require some combination of the following inputs: a network consisting of nodes and links, and a travel time function on each link, which could be a time-independent or a time-dependent function, where the time-dependent functions account for time-of-day delays resulting from actual travel conditions such as peak-hour congestion. The problem then seeks a shortest path between one or more origin-destination pairs. A new variant, introduced in the context of TRANSIMS and which is the focus of the present study, also specifies labels for each arc denoting particular modes of travel, along with strings of admissible labels that delineate the permissible travel mode sequences that could be adopted by the user in traveling from the origin to the destination of the trip. The technique adopted by TRANSIMS to identify a suitable travel route for any user is a variant of Dijkstra's procedure for finding shortest paths, which is suitably modified to accommodate time-dependent travel times and label sequence constraints. The underlying problem is referred to as a Time-Dependent Label-Constrained Shortest Path Problem. The main objective of this research is to improve upon this procedure and study its implementation in order to develop a more effective scheme for determining time-dependent label-constrained shortest paths as a practical routing tool in multimodal transportation networks. Specifically, we enhance the following features of this procedure: (a) We recommend a method to work implicitly with a certain composition graph G* that combines the transportation network with the admissible label-sequence graph. This graph G* captures all possible paths for a given single trip starting from the origin node and ending at the destination node, while conforming with the admissible mode string. (b) We use more modern partitioned shortest path algorithmic schemes to implement the time-dependent label-constrained procedure. (c) We introduce the notion of curtailing search based on various indicators of progress and projected travel times to complete the trip. Finally, computer programs in C++ are written to implement the proposed overall algorithm, and are applied to solve some real multimodal transportation network problems. The indicators used to evaluate the performance of the algorithm include (i) time taken for computation on the real network, (ii) quality of solution obtained, (iii) ease of implementation, and (iv) extensibility of the algorithm for solving other variants of the shortest path problem. The results exhibit that the proposed algorithm, even without the approximate curtailing of the search process, exhibits good performance in finding optimal routes for real multimodal transportation networks. Although the various heuristic curtailments result in only approximate solutions, typically, they run much faster than the exact algorithm for the intuitive reason that the shortest path tree developed grows more pointedly in the direction of the destination. Among the different strategies implemented, our results suggest that the scheme based on the geometric structure of the underlying network, using either a constant predictive term, or multiplying this term with a suitable exponential decay function, yields an attractive candidate for heuristically curtailing the search. / Master of Science

This thesis is divided into two parts that each is concerned with a specific problem. The problem under consideration in the first part is to find suitable graph representations, abstractions, cost measures and algorithms for calculating placements of unmanned aerial vehicles (UAVs) such that they can keep one or several static targets under constant surveillance. Each target is kept under surveillance by a surveillance UAV, which transmits information, typically real time video, to a relay UAV. The role of the relay UAV is to retransmit the information to another relay UAV, which retransmits it again to yet another UAV. This chain of retransmission continues until the information eventually reaches an operator at a base station. When there is a single target, then all Pareto-optimal solutions, i.e. all relevant compromises between quality and the number of UAVs required, can be found using an efficient new algorithm. If there are several targets, the problem becomes a variant of the Steiner tree problem and to solve this problem we adapt an existing algorithm to find an initial tree. Once it is found, we can further improve it using a new algorithm presentedin this thesis. The second problem is optimization of time-consuming problems where the objective function is seen as a black box, where the input parameters are sent and a function valueis returned. This has the important implication that no gradient or Hessian information is available. Such problems are common when simulators are used to perform advanced calculations such as crash test simulations of cars, dynamic multibody simulations etc. It is common that a single function evaluation takes several hours. Algorithms for solving such problems can be broadly divided into direct search algorithms and model building algorithms. The first kind evaluates the objective function directly, whereas the second kind builds a model of the objective function, which is then optimized in order to find a new point where it is believed that objective function has agood value. Then the objective function is evaluated in that point. Since the objective function is very time-consuming, it is common to focus on minimizing the number of function evaluations. However, this completely disregards the possibility to perform calculations in parallel and to exploit this we investigate different ways parallelization can be used in model-building algorithms. Some of the ways to do this is to use several starting points, generate several new points in each iteration, new ways of predicting a point’s value and more. We have implemented the parallel extensions in one of the state of the art algorithms for derivative-free optimization and report results from testing on synthetic benchmarksas well as from solving real industrial problems.

abstract: Vehicles powered by electricity and alternative-fuels are becoming a more popular form of transportation since they have less of an environmental impact than standard gasoline vehicles. Unfortunately, their success is currently inhibited by the sparseness of locations where the vehicles can refuel as well as the fact that many of the vehicles have a range that is less than those powered by gasoline. These factors together create a "range anxiety" in drivers, which causes the drivers to worry about the utility of alternative-fuel and electric vehicles and makes them less likely to purchase these vehicles. For the new vehicle technologies to thrive it is critical that range anxiety is minimized and performance is increased as much as possible through proper routing and scheduling. In the case of long distance trips taken by individual vehicles, the routes must be chosen such that the vehicles take the shortest routes while not running out of fuel on the trip. When many vehicles are to be routed during the day, if the refueling stations have limited capacity then care must be taken to avoid having too many vehicles arrive at the stations at any time. If the vehicles that will need to be routed in the future are unknown then this problem is stochastic. For fleets of vehicles serving scheduled operations, switching to alternative-fuels requires ensuring the schedules do not cause the vehicles to run out of fuel. This is especially problematic since the locations where the vehicles may refuel are limited due to the technology being new. This dissertation covers three related optimization problems: routing a single electric or alternative-fuel vehicle on a long distance trip, routing many electric vehicles in a network where the stations have limited capacity and the arrivals into the system are stochastic, and scheduling fleets of electric or alternative-fuel vehicles with limited locations to refuel. Different algorithms are proposed to solve each of the three problems, of which some are exact and some are heuristic. The algorithms are tested on both random data and data relating to the State of Arizona. / Dissertation/Thesis / Ph.D. Industrial Engineering 2014

The Constrained Shortest Path Problem (CSPP) consists of finding the shortest path in a graph or network that satisfies one or more resource constraints. Without these constraints, the shortest path problem can be solved in polynomial time; with them, the CSPP is NP-hard and thus far no polynomial-time algorithms exist for solving it optimally. The problem arises in a number of practical situations. In the case of vehicle path planning, the vehicle may be an aircraft flying through a region with obstacles such as mountains or radar detectors, with an upper bound on the fuel consumption, the travel time or the risk of attack. The vehicle may be a submarine travelling through a region with sonar detectors, with a time or risk budget. These problems all involve a network which is a discrete model of the physical domain. Another example would be the routing of voice and data information in a communications network such as a mobile phone network, where the constraints may include maximum call delays or relay node capacities. This is a problem of current economic importance, and one for which time-sensitive solutions are not always available, especially if the networks are large. We consider the simplest form of the problem, large grid networks with a single side constraint, which have been studied in the literature. This thesis explores the application of Constraint Programming combined with Lagrange Relaxation to achieve optimal or near-optimal solutions of the CSPP. The following is a brief outline of the contribution of this thesis. Lagrange Relaxation may or may not achieve optimal or near-optimal results on its own. Often, large duality gaps are present. We make a simple modification to Dijkstra’s algorithm that does not involve any additional computational work in order to generate an estimate of path time at every node. / We then use this information to constrain the network along a bisecting meridian. The combination of Lagrange Relaxation (LR) and a heuristic for filtering along the meridian provide an aggressive method for finding near-optimal solutions in a short time. Two network problems are studied in this work. The first is a Submarine Transit Path problem in which the transit field contains four sonar detectors at known locations, each with the same detection profile. The side constraint is the total transit time, with the submarine capable of 2 speeds. For the single-speed case, the initial LR duality gap may be as high as 30%. The first hybrid method uses a single centre meridian to constrain the network based on the unused time resource, and is able to produce solutions that are generally within 1% of optimal and always below 3%. Using the computation time for the initial Lagrange Relaxation as a baseline, the average computation time for the first hybrid method is about 30% to 50% higher, and the worst case CPU times are 2 to 4 times higher. The second problem is a random valued network from the literature. Edge costs, times, and lengths are uniform, randomly generated integers in a given range. Since the values given in the literature problems do not yield problems with a high duality gap, the values are varied and from a population of approximately 100,000 problems only the worst 200 from each set are chosen for study. These problems have an initial LR duality gap as high as 40%. A second hybrid method is developed, using values for the unused time resource and the lower bound values computed by Dijkstra’s algorithm as part of the LR method. The computed values are then used to position multiple constraining meridians in order to allow LR to find better solutions. / This second hybrid method is able to produce solutions that are generally within 0.1% of optimal, with computation times that are on average 2 times the initial Lagrange Relaxation time, and in the worst case only about 5 times higher. The best method for solving the Constrained Shortest Path Problem reported in the literature thus far is the LRE-A method of Carlyle et al. (2007), which uses Lagrange Relaxation for preprocessing followed by a bounded search using aggregate constraints. We replace Lagrange Relaxation with the second hybrid method and show that optimal solutions are produced for both network problems with computation times that are between one and two orders of magnitude faster than LRE-A. In addition, these hybrid methods combined with the bounded search are up to 2 orders of magnitude faster than the commercial CPlex package using a straightforward MILP formulation of the problem. Finally, the second hybrid method is used as a preprocessing step on both network problems, prior to running CPlex. This preprocessing reduces the network size sufficiently to allow CPlex to solve all cases to optimality up to 3 orders of magnitude faster than without this preprocessing, and up to an order of magnitude faster than using Lagrange Relaxation for preprocessing. Chapter 1 provides a review of the thesis and some terminology used. Chapter 2 reviews previous approaches to the CSPP, in particular the two current best methods. Chapter 3 applies Lagrange Relaxation to the Submarine Transit Path problem with 2 speeds, to provide a baseline for comparison. The problem is reduced to a single speed, which demonstrates the large duality gap problem possible with Lagrange Relaxation, and the first hybrid method is introduced. / Chapter 4 examines a grid network problem using randomly generated edge costs and weights, and introduces the second hybrid method. Chapter 5 then applies the second hybrid method to both network problems as a preprocessing step, using both CPlex and a bounded search method from the literature to solve to optimality. The conclusion of this thesis and directions for future work are discussed in Chapter 6.

Cette thèse développe des algorithmes pour les problèmes de plus court chemin sous cont-rain-tes de ressources, et les applique à l'optimisation des rotations des avions et des équipages d'une compagnie aérienne dans le cadre d'approches par génération de colonnes.Les problèmes de plus court chemin sous contraintes de ressources sont généralement résolus grâce à une énumération intelligente de tous les chemins non dominés. Les approches récentes utilisent des bornes sur les ressources des chemins pour éliminer des solutions partielles. L'efficacité de la méthode est conditionnée par la qualité des bornes utilisées. Notre principale contribution au domaine est l'introduction d'une procédure générique pour calculer des bornes qui s'applique à la plupart des problèmes de chemins sous contraintes, et en particulier les problèmes stochastiques. A cette fin, nous introduisons une généralisation du problème de plus court chemin sous contraintes dans laquelle les ressources des chemins appartiennent à un monoïde ordonné comme un treillis. La ressource d'un chemin est la somme des ressources de ses arcs, le terme somme désignant l'opérateur du monoïde. Le problème consiste à trouver parmi les chemins qui satisfont une contrainte donnée celui dont la ressource minimise une fonction de coût croissante de la ressource des chemins. Nous généralisons les algorithmes d'énumération à ce nouveau problème. La théorie des treillis nous permet de construire une procédure polynomiale pour trouver des bornes de qualité. L'efficacité pratique de la méthode est évaluée au travers d'une étude numérique détaillée sur des problèmes de chemins déterministes et stochastiques. Les procédures de calcul des bornes peuvent être interprétées comme des généralisations aux monoïdes ordonnés comme des treillis d'algorithmes de la littérature définis pour résoudre un problème de chemin pour lequel les ressources des chemins prennent leur valeur dans un semi-anneau.Nos algorithmes de chemins ont été appliqués avec succès au problème de crew pairing. Étant donné un ensemble de vols opérés par une compagnie aérienne, les problèmes d'aircraft routing et de crew pairing construisent respectivement les séquences de vols opérées par les avions et par les équipages de manière à couvrir tous les vols à moindre coût. Comme certaines séquences de vols ne peuvent être réalisées par un équipage que s'il reste dans le même avion, les deux problèmes sont liés. La pratique actuelle dans l'industrie aéronautique est de résoudre tout d'abord le problème d'aircraft routing, puis le problème de crew pairing, ce qui aboutit à une solution non-optimale. Des méthodes de résolution pour le problème intégré ont été développées ces dix dernières années. Nous proposons une méthode de résolution pour le problème intégré reposant sur deux nouveaux ingrédients : un programme linéaire en nombre entier compact pour le problème d'aircraft routing, ainsi que de nouveaux pour le problème esclave de l'approche usuelle par génération de colonnes du problème de crew pairing. Ces algorithmes pour le problème esclave sont une application de nos algorithmes pour le problème de plus court chemin sous contraintes. Nous généralisons ensuite cette approche de manière à prendre en compte des contraintes de probabilités sur la propagation du retard. Ces algorithmes permettent de résoudre quasiment à l'optimum les instances industrielles d'Air France / This thesis develops algorithms for resource constrained shortest path problems, and uses them to solve the pricing subproblems of column generation approaches to some airline operations problems.Resource constrained shortest path problems are usually solved using a smart enumeration of the non-dominated paths. Recent improvements of these enumeration algorithms rely on the use of bounds on path resources to discard partial solutions. The quality of the bounds determines the performance of the algorithm. Our main contribution to the topic is to introduce a standard procedure to generate bounds on paths resources in a general setting which covers most resource constrained shortest path problems, among which stochastic versions. In that purpose, we introduce a generalization of the resource constrained shortest path problem where the resources are taken in a lattice ordered monoid. The resource of a path is the monoid sum of the resources of its arcs. The problem consists in finding a path whose resource minimizes a non-decreasing cost function of the path resource among the paths that satisfy a given constraint. Enumeration algorithms are generalized to this framework. We use lattice theory to provide polynomial procedures to find good quality bounds. The efficiency of the approach is proved through an extensive numerical study on deterministic and stochastic path problems. Interestingly, the bounding procedures can be seen as generalizations to lattice ordered monoids of some algebraic path problem algorithms which initially work with resources in a semiring.Given a set of flight legs operated by an airline, the aircraft routing and the crew pairing problem build respectively the sequences of flight legs operated by airplanes and crews at minimum cost. As some sequences of flight legs can be operated by crews only if they stay in the same aircraft, the two problems are linked. The current practice in the industry is to solve first the aircraft routing, and then the crew pairing problem, leading to a non-optimal solution. During the last decade, solution schemes for the integrated problem have been developed. We propose a solution scheme for the integrated problem based on two new ingredients: a compact integer program approach to the aircraft routing problem, and a new algorithm for the pricing subproblem of the usual column generation approach to the crew pairing problem, which is based on our resource constrained shortest path framework. We then generalize the algorithm to take into account delay propagation through probabilistic constraints. The algorithms enable to solve to near optimality Air France industrial instances