This dissertation is devoted to the modeling, analysis and development of solution approaches for some optimization-related problems encountered in industrial and manufacturing settings. We begin by introducing a special type of traveling salesman problem called "High Multiplicity Asymmetric Traveling Salesman Problem" (HMATSP). We propose a new formulation for this problem, which embraces a flow-based subtour elimination structure, and establish its validity for this problem. The model is, then, incorporated as a substructure in our formulation for a lot-sizing problem involving parallel machines and sequence-dependent setup costs, also known as the "Chesapeake Problem". Computational results are presented to demonstrate the efficacy of our modeling approach for both the generic HMATSP and its application within the context of the Chesapeake Problem. Next, we investigate an integrated lot-sizing and scheduling problem that is encountered in the primary manufacturing facility of pharmaceutical manufacturing. This problem entails determination of production lot sizes of multiple products and sequence in which to process the products on machines, which can process lots (batches) of a fixed size (due to limited capacity of containers) in the presence of sequence-dependent setup times/costs. We approach this problem via a two-stage optimization procedure. The lot-sizing decision is considered at stage 1 followed by the sequencing of production lots at stage 2. Our aim for the stage 1 problem is to allocate batches of products to time-periods in order to minimize the sum of the inventory and backordering costs subject to the available capacity in each period. The consideration of batches of final products, in addition to those for intermediate products, which comprise a final product, further complicates the lot-sizing problem. The objective for the stage 2 problem is to minimize sequence-dependent setup costs. We present a novel unifying model and a column generation-based optimization approach for this class of lot-sizing and sequencing problems. Computational experience is first provided by using randomly generated data sets to test the performances of several variants of our proposed approach. The efficacy of the best of these variants is further demonstrated by applying it to the real-life data collected with the collaboration of a pharmaceutical manufacturing company. Then, we address a single-lot, lot streaming problem for a two-stage assembly system. This assembly system is different from the traditional flow shop configuration. It consists of m parallel subassembly machines at stage 1, each of which is devoted to the production of a component. A single assembly machine at stage 2, then, assembles products after components (one each from the subassembly machines at the first stage) have been completed. Lot-detached setups are encountered on the machines at the first and second stages. Given a fixed number of transfer batches (or sublots) from each of the subassembly machines at stage 1 to the assembly machine at stage 2, our problem is to find sublot sizes so as to minimize the makespan. We develop optimality conditions to determine sublot sizes for the general problem, and present polynomial-time algorithms to determine optimal sublot sizes for the assembly system with two and three subassembly machines at stage 1. Finally, we extend the above single-lot, lot streaming problem for the two-stage assembly system to multiple lots, but still, for the objective of minimizing the makespan. Due to the presence of multiple lots, we need to address the issue of the sequencing of the lots along with lot-splitting, a fact which adds complexity to the problem. Some results derived for the single-lot version of this problem have successfully been generalized for this case. We develop a branch-and-bound-based methodology for this problem. It relies on effective lower bounds and dominance properties, which are also derived. Finally, we present results of computational experimentation to demonstrate the effectiveness of our branch-and-bound-based methodology. Because of the tightness of our upper and lower bounds, a vast majority of the problems can be solved to optimality at root node itself, while for others, the average gap between the upper and lower bounds computed at node zero is within 0.0001%. For a majority of these problems, our dominance properties, then, effectively truncate the branch-and-bound tree, and obtain optimal solution within 500 seconds. / Ph. D.

Die gängige Methodik zum Konzipieren automatisierter Montagesysteme umfasst die Planung der Montage und die physische Entwicklung der technischen Gesamtlösung. Zur Abbildung der konkreten Aufgabe werden standardisierte Symbole in einem Ablaufplan miteinander verknüpft. Die Hauptaufgabe des Konstrukteurs ist die Auswahl und die Zusammenstellung einer optimalen Konfiguration der Funktionsträger sowie deren Implementierung in einer Gesamtlösung unter Berücksichtigung der vorgegebenen Randbedingungen. Das Problem ist der fehlende Informationsgehalt der bisher eingesetzten Handhabungssymbole und Symbole zur Ermittlung der Funktionsträger zur Beschreibung der Montage- und Handhabungsplanung sowie die fehlende Vorgehensweise zur Auswahl der Funktionsträger aus den verschiedenen Varianten nach minimaler Taktzeit und Gesamtanschaffungskosten. Zur Realisierung eines effizienten Montageplanungssystems leitet sich daher die Zielstellung ab, den Informationsgehalt der standardisierten Symbole zu erweitern und mit logischen Schnittstellen für eine automatisierte Verknüpfung in der Funktionsfolge auszustatten. Diese neuen Symbole beinhalten die Definition der Funktionen sowie alle Randbedingungen und Parameter zur eindeutigen Beschreibung der Handhabungsaufgabe. Mithilfe dieser Parameter werden Anforderungslisten erstellt und nach passenden Anlagenkomponenten gesucht. Zur Auswahl der optimalen Komponenten des Montagesystems wird das lineare Optimierungsproblem hinsichtlich der Kombination aus Taktzeit und Gesamtanschaffungskosten gelöst. / The common methodology for designing automated assembly systems involves the assembly planning and the physical development of overall technical solution. To illustrate the concrete task, standardized symbols are connected together in a flowchart. The designer's main task is the selection and the composition of an optimal configuration of the functional carriers as well as their implementation in an overall solution in consideration of the predetermined boundary conditions. One problem is the lack of information content of the previously used handling symbols and the symbols for determining the functional carriers, which describe the assembly and handling planning. The other is the insufficient methods for selecting the functional carriers from the different variants based on minimum cycle time and total acquisition cost. In order to realize an efficient assembly planning system, the objective is therefore to expand the information content of the standardized symbols and equip them with logical interfaces for automated connection in the functional sequence. These new symbols contain the definition of the functions as well as all boundary conditions and parameters for the unambiguous description of the handling task. These parameters are utilised to create requirement lists and search for suitable plant components. In order to select the optimal components of the assembly system, the linear optimization problem regarding the combination of cycle time and total acquisition costs is solved.

Examensarbetet utfördes våren 2018 på Scanias axel- och växellådsmontering i Södertälje. Axel- och växellådsmonteringen består till största del av manuella monteringsstationer men också till en del av automationslösningar för skruvdragning. Huvudmålet för examensarbetet är att så långt som möjligt automatisera tre manuella monteringsstationer, med delmålen att redovisa nuläget med eventuellt nödvändiga förändringar i processen samt identifiera tekniska lösningar för automation. De tekniska lösningarna för automation ska vara nedbrutna i en budget och payback för eventuell investering. Mål och avgränsningar för rapporten återfinns i inledningskapitlet. För att få operations- och produktkännedom inleds examensarbetet med en praktik på de berörda monteringsstationerna. Vidare fortsatte arbetet med litteraturstudier för att ge en teoretisk referensram. Den teoretiska referensramen används dels i rapporten samt även som underlag för ett flertal intervjuer med forskare på KTH och Scania, samt med kollegor inom produktionsteknik. För att finna tekniklösningar utfördes intern benchmark på flera av Scanias produktionsenheter. Intern benchmark och litteraturstudie har varit grunden för examensarbetets benchmark. I kapitel 2.1 återfinns en utförlig beskrivning av examensarbetets arbetsmetod samt de metoder som tillämpats. De monteringsmoment som utförs på den line där de manuella monteringsstationerna är placerade är att sammanfoga skruv-, press-, limförband. En beskrivning av linen och vilka monteringsmoment som utförs på vardera station återfinns i kapitel 4.2, figur 9. De monteringsoperationerna som utförs på de manuella monteringsstationerna är skruvilägg, mutteräntring, limning, dragning, lagerpressning och avlyft. Examensarbetet visar i kapitel 6.2 de tekniklösningar som kan användas för automation av de fyra manuella monteringsstationerna. Tekniklösningarna är identifierade under intern benchmark på Scania. Tekniklösningarna bryts ned i investeringsuppskattningar vilka presenteras i kapitel 6.3, tabell 15. Tekniklösningarna uppskattas att för hela systemet kräva en investering på 16.5 till 22.5 miljoner kronor. En investering i tekniklösningarna ger då en årlig besparing om fem positioner vilket resulterar i en årlig besparing om 2.5 miljoner kronor. Payback-tid för investeringen (kapitel 6.3, tabell 16) blir då 6.6 till 9 år. Fördelar med att automatisera de manuella monteringsstationerna blir att de ergonomiska repetitionsbelastningarna för operatörerna minskar (kapitel 3.4 och 4.4) samt att kvalitetssäkring av limförband kan utföras vilket i nuläget är utmanande. Utmaningarna med att automatisera blir att skapa värde för fler företagsdelar i samband med automationsinvesteringen samt skapa en flexibel automationslösning vilken kan möta framtida produkt- eller produktionsvolymsförändringar. En rekommendation i att investera i examensarbetets framställda tekniklösningar för automation kan inte ges, däremot rekommenderas att studera vidare hur ett alternativ skulle se ut för att vara flexibelt nog att möta framtida förändringar samt skapa för flera delar av företaget vid investering. / The thesis project is conducted during spring of 2018 at Scania axle and transmission assembly in Södertälje. The axle and transmission assembly consists in large parts of manual assembly stations but also of automatic tightening solution The main objective of the thesis project is to automate three manual assembly stations as far as possible. The main objective where divided into three assignments. First assignment is to identify and describe the work stations current state. Second assignment is to identify and propose necessary changes of the current process. Last assignment is to identify technical solutions for automation and present investment, budget and calculate time for payback. Objectives and delimitations of the thesis project is found in the introduction chapter. To gather knowledge of the product and assembly line an practice period where conducted during the first week of the thesis project. To reach an academic perspective of our thesis a literature review was conducted. The theoretical frame of reference is used during the project in two different ways, as a theoretical background to our objective and also as an basis for several interviews with researchers and senior colleagues at Scania. In order to meet the third assignment to identify technical solutions an internal benchmark was performed on several of Scanias production units. Internal benchmark and literature review has been the basis of the proposed technical solutions. Chapter 2.1 contains a description of work method of the thesis and the methods applied. The manual assembly operations performed on the assembly line are to assemble screw, press and adhesives conjoins. A description of the manual assembly operations is found in chapter 4.2, Figure 9. The assembly operations performed on the manual assembly stations are screw and nut placement, adhesive application, bearing placement and lift. The thesis describes in chapter 6.2 the technical options for automation solutions applied to the four manual assembly stations. The technical options are identified through the internal benchmark at Scania. The technical options are presented as estimated investments in Chapter 6.3, table 15. The technical option is estimated to an investment of 16.5 to 22.5 million sek. The technical solution results in an annual saving at 2.5 million sek., which gives an 6.6 to 9 years payback time (Shown in Chapter 6.3, table 16). The automation solution the thesis project proposes has advantages in that the ergonomic repetitions are significantly reduced (chapter 3.4 and 4.4) and contributes to ensure quality of the adhesive application of the conjoint. Challenges with the proposed automation system is to create additional value cross functional through the organization and to design a solution that is flexible enough to meet changes in products and/or production flow. Recommendation for investment in the thesis projects automation solution cannot be given due to the lack of future flexibility and additional value to another part of the organization. The thesis project shows significant advantages for automation solution in terms of ergonomics and quality. The recommendation is to continue the study of how an investment in an automated solution can create a cross functional value and contribute to product flexibility.

The forming of an assembly system is a complex task, which should be considered as never ending. In order to successfully plan and implement an assembly system it is of vital importance that the obstacles and preconditions that have an impact on the system are identified and evaluated. This together with the necessary support activities and the attributes of the product to be assembled constitutes the starting point for the forming of the assembly system. The aim of this thesis is to link the theoretical findings with the issues stated above, and through this explain a best practice approach when forming the assembly system. The theoretical work aims at describing the nature and activities within assembly and manufacturing systems and explains these in three different levels of strategies divided into Manufacturing strategies, Layout, material flow and design strategies and finally Logistic, material handling and quality strategies. Then the obstacles and preconditions found are discussed and evaluated which set the basis for the forming of the assembly system and by linking these with the relevant theory, conceptual design proposals for the assembly system and the Logistic support system are formed. These are then evaluated and finally a proposal for the detailed layout of the assembly system is given. This proposal is then to be used as a guideline for the company Aercrete when forming their assembly system.

<p>The forming of an assembly system is a complex task, which should be considered as never ending. In order to successfully plan and implement an assembly system it is of vital importance that the obstacles and preconditions that have an impact on the system are identified and evaluated. This together with the necessary support activities and the attributes of the product to be assembled constitutes the starting point for the forming of the assembly system.</p><p>The aim of this thesis is to link the theoretical findings with the issues stated above, and through this explain a best practice approach when forming the assembly system. The theoretical work aims at describing the nature and activities within assembly and manufacturing systems and explains these in three different levels of strategies divided into Manufacturing strategies, Layout, material flow and design strategies and finally Logistic, material handling and quality strategies. Then the obstacles and preconditions found are discussed and evaluated which set the basis for the forming of the assembly system and by linking these with the relevant theory, conceptual design proposals for the assembly system and the Logistic support system are formed.</p><p>These are then evaluated and finally a proposal for the detailed layout of the assembly system is given. This proposal is then to be used as a guideline for the company Aercrete when forming their assembly system.</p>