Lagerhantering anses vara en betydelsefull aktivitet som företag måste ta sig an, då förbättrade metoder kan ha direkt påverkan på resultatet. Produktallokering inom lagret är en av de vitala delarna inom lagerplanering. Detta arbete är en fallstudie där fallet är ett lager hos ett producerande företag med det övergripande syftet att undersöka hur en optimeringsmodell för lagerhantering kan hjälpa organisationer med dess lagerhantering genom att studera rådande processer och utifrån det skapa en optimeringsmodell för att beräkna optimal lagerallokering. Studien har utförts genom en mixad metod, uppdelad i en kvantitativ- och en kvalitativ fas, då forskning hävdar att nyckeln till framgång för effektiv lagerhantering är att matcha policyn med lagrets layout och de ordertyper som lagret hanterar, för att anpassa optimeringsmodellen utifrån det givna fallets policy, lagerlayout och ordertyper. I den första fasen genomfördes en processkartläggning för att bygga upp för optimeringens utformning. Den andra fasen gick ut om att matematiskt beskriva samma problemuppställning och modellera problemet på insamlad kvantitativ data. I processkartläggningsfasen identifierades de två differentieringskriterierna omsättning i enheten sålda paket och förädling i enheten produktionskostnad. Målet med utförd klassificering var att ta fram vilka unika produkter som bör behandlas differentierat och erhålla tilldelad lagring, vilket var en förbättringsmöjlighet funnen i nulägesbeskrivningen. Totalt togs det fram 93 produkter att placera ut på någon av de 220 lagringsplatserna med målet att minimera körsträckan med truck. Jämförelse av målfunktionsvärden mellan optimaluppsättning enlig modell och dagens uppsättning, påvisade en teoretisk besparing av antal körda meter på 35 %. Studiens resultat har påvisat att matematisk modellering och linjärprogrammering går att använda för lagerallokeringsproblem för att optimera produktallokeringen, där den matematiskmodellen baserats på fallspecifika egenskaper. Modellens resultat förväntas underlätta ett antal av de problem som noterats vid kartläggning av lagerhanteringsprocesserna. / Inventory management is considered to be an important activity that companies need to accomplish as improved methods directly can affect the final result. Product allocation within the warehouse is one of the vital parts of inventory management. This report is a case study where a warehouse of a manufacturing company is considered the case. The overall purpose is to investigate how a optimization model can help organizations with their inventory management, by first studying existing processes and based on that create the model to calculate optimal inventory allocation. The study has been conducted through a mixed method, divided into a quantitative and qualitative phase, as research claims that the key to the success of efficient inventory management is to match the layout of the warehouse case policy, stock layout and order types in order to customize the optimization model. In the first phase, a process mapping was conducted to build the optimization design. The second phase was based on mathematically describing the same problem setup, and then to model the problem with collected data. In the process-mapping phase, two differentiation criteria were identified as units sold packages and the unit production cost. The purpose of the classification was to identify which unique products that should be treated differently and receive assigned storage, which was an improvement opportunity found in the case description. A total of 93 products were placed on one of the 220 storage sites, with the aim of minimizing the mileage by truck. Comparison of target function values ​​between the optimal set-up model and the current set, demonstrated a theoretical saving of the number of driven meters of 35%. The study results have shown that mathematical modeling and linear programming can be used for inventory allocation problems to optimize product allocation, where the mathematical model is based on case-specific properties. The model's results are expected to facilitate a number of the problems noted in mapping inventory management processes.

Epiroc har åtnjutit en längre period av kraftig tillväxt samtidigt som flertalet effektiviseringsprojekt har genomförts med syfte att öka produktionsvolymen för att möta marknadens efterfrågan. Detta genom att bland annat implementera en variant av Lean production – The Way We Produce. En av dessa förändringar är en takad flödesorienterad montering med just-in-time sekvenserad materialförsörjning. Detta har i sin tur ökat kraven på materialförrådet Logistikcenter (LC) där ledningen nu börjat undersöka möjliga effektiviseringsåtgärder. Därav är studiens syfte till att ta fram realiserbara förbättringsförslag för logistikverksamheten vid Logistikcenter tillhörande Epiroc Rock Drills AB i Örebro för att öka effektiviteten och leveranssäkerheten. Studien har genomförts i två faser; identifieringsfasen och fördjupningsfasen. Under identifieringsfasen genomfördes en kartläggning av nuläget i LC, där underlaget för kartläggningen baserar sig på intervjuer, observationer och analyser. Genom en rotorsaksanalys, med målet att identifiera källor till ineffektivitet, kunde nio förbättringsområden identifieras varav en av dessa vidare skulle utredas i fördjupningsfasen. Dessa utvärderades utifrån en effekt-insats matris för att välja det förbättringsområde med störst effektiviseringspotential i förhållande till den förväntande insatsen. Analysmodellen för effekt-insats matrisen var de åtta slöserierna i Lean (Petersson, et al., 2015), dess förväntade påverkan på effektiviteten samt den förväntade komplexiteten av en implementation. Resultatet från rotorsaksanalysen gav artikelklassificering som det primära förbättringsområdet där den undersökta åtgärden var en alternativ tillämpning av frekvensläggning för att minimera rörelsetiden mellan lagerplatser vid plock. Under fördjupningsfasen undersöktes den nuvarande artikelklassificeringen och frekvensläggningen genom syntes av en alternativ modell för klassificeringen av artiklar och lagerplatser utifrån en fördjupad litteraturstudie. För att undersöka om en alternativ klassificering kan öka effektiviteten utvecklades en utvärderingsmodell som modellerar rörelsetiden för historiskt data från plocklistor. Den användes för att testa vilken kombination av storlekar på artikelklasserna som gav den minsta möjliga totala rörelsetiden. En kombination av 60/30/10 % (A/B/C) av det ackumulerade antalet plock gav den lägsta totala rörelsetiden för en plockhistorik på 15 månader med en reduktion av rörelsetiden motsvarande 33 % (1760 h) jämfört mot nuläget. Vidare undersöktes en alternativ sortering av plocklistorna för automathissar respektive pallställage med utvärderingsmodellen som gav en reduktion på 4 % respektive 11 % mot nuläget. Slutligen undersöktes olika former på zonerna för klassificering av lagerplatser i pallställage W3, där utlämningsplatsen är placerad halvvägs in i ställaget. En tyngdpunkt placerat centralt mellan ingången och utlämningsplatsen gav det bästa simuleringsresultatet. Sammanfattningsvis fastställdes artikelklassificering som det förbättringsområde med störst realiserbar effektiviseringspotential med lägst komplexitet utifrån en rotsorsaksanalys. Epiroc rekommenderas att implementera klassificeringsmodellen med tre klasser av storleken 60/30/10 % (A/B/C) av den ackumulerade antalet plockrader, samt att implementera att den alternativa sorteringen av plockrader på plocklistor för pallställage. / Epiroc has had a long period of strong growth, where several efficiency projects have been implemented to increase production volume in order to meet market demand. A variant of Lean production has been implemented over several years – called The Way We Produce by Epiroc. A large part of that change has been the implementation of sequenced flow-oriented assembly with just-in-time sequenced material deliveries to the assembly floor. This, in turn, has increased the requirements for the warehouse Logistic Center (LC) and management has now begun to investigate ways to improve efficiency at LC. Hence, the aim of the study is to develop realistic improvement proposals for the logistics operations at Logistics Center of Epiroc Rock Drills AB at Örebro to increase efficiency and delivery reliability. The study has been conducted in two phases; the identification phase and the in-depth phase. During the identification phase, a mapping of the current situation in LC was carried out, where the basis for the survey is based on interviews, observations and analyzes. Through a root cause analysis, with the goal of finding sources of inefficiency, nine areas of improvement could be identified, one of which would be chosen for further investigation in the in-depth phase. The areas of improvement were evaluated with an effect-input matrix to choose the area of improvement that provides the greatest efficiency potential in relation to the expected effort. The analysis model for the effect-input matrix was the eight wastes of Lean based on Petersson et al. (2015), its expected impact on efficiency and the expected complexity of an implementation. The result of the root cause analysis gave article classification as the primary area of improvement, where the measure is an alternative slotting strategy to minimize the movement time between storage locations. During the in-depth phase, the current article classification and frequency setting were examined by synthesis of an alternative model for the classification of articles and storage locations based on an in-depth literature study. To investigate whether an alternative classification can increase efficiency, an evaluation model was developed that models the movement time from historical data from pick lists. It was used to test which combination of sizes for the article classes gave the smallest possible total movement time. A combination of 60/30/10% (A/B/C) gave the lowest total movement time for a picking history of 15 months with a reduction corresponding to 33% (1760 h) compared to the current situation. Furthermore, an alternative sorting of the pick lists for the vertical lift modules and pallet racking was examined with the evaluation model, which gave a reduction of 4% and 11% respectively. Finally, various forms for the zones were examined for the classification of storage locations in pallet rack W3, where the delivery site is located three-quarter way into the pallet rack. A center of gravity for the for the A-class placed between the entrance and the delivery point gave the best simulation result. In summary, the article classification improvement area was established with the greatest realizable efficiency potential with the least effort based on the root cause analysis. Epiroc is recommended to implement the classification model with three classes of size 60/30/10 % (A/B/C) of the accumulated number of picking rows, and to implement that alternative sorting of pick rows on picklists for pallet racking.

La presente tesis tiene como objetivo identificar si existe correlación alguna entre los costos evitables en operaciones de importación marítima y la prestación de nivel de servicio mostrando resultados sobre la implementación realizada en una de las principales empresas en el rubro de importación y distribución de equipamiento gastronómico profesional de la ciudad de Lima. Luego de realizar la implementación aprobada para el 2017, los costos evitables en operaciones de importación marítima fueron reducidos en un 13.85% con respecto al periodo anterior. Se demostró que los costos evitables (r=0.75) tiene una relación alta y positiva con el nivel de servicio, es decir que a medida que se disminuyan y la empresa deje de asumir estos podrá ser mejor su nivel de servicio. Para la presentación se tiene una estructura de cinco capítulos: En el primer capítulo, el marco teórico, se detalló conceptos relacionados a los costos evitables en operación de importación marítima, nivel de servicio, cantidad óptima a ordenar, inventario de seguridad, punto de reorden y costos totales de gestión de inventarios. Además, se mostró información sobre la empresa donde se implementaron las mejoras. En el segundo capítulo, se explicó temas en relación al plan de investigación, dando a conocer el problema, la formulación de la hipótesis, el objetivo general y los específicos. En el tercer capítulo, se trató la metodología de la investigación, donde se determinó el enfoque y diseño, la población, la definición de variables y la recolección de datos. En el cuarto capítulo, se desarrolló la investigación calculando los costos evitables en 3 niveles de servicio. Se estableció dos escenarios (con costos evitables y sin costos evitables) a los sub-variables niveles de servicio, cantidad óptima a ordenar, inventario de seguridad, punto de reorden y costos totales de gestión de inventarios. Finalmente, en el quinto capítulo se mostró el análisis de resultados respondiendo al problema si existe correlación entre los costos evitables en operación de importación marítima y la prestación de nivel de servicio, y brindaremos conclusiones y las recomendaciones sobre la investigación desarrollada. / The objective of this thesis is to identify if there is any correlation between the avoidable costs in maritime import operations and the level of service provision, showing results on the implementation carried out in one of the main companies in the import and distribution of professional gastronomic equipment category from the city of Lima. After implementing the approved implementation for 2017, avoidable costs in maritime import operations were reduced by 13.85% compared to the previous period. It was shown that the avoidable costs (r = 0.75) has a high and positive relationship with the service level, that is to say that as they decrease and the company stops assuming these, their level of service may be better. For the presentation there is a structure of five chapters: In the first chapter, the theoretical framework, detailed concepts related to avoidable costs in maritime import operation, service level, optimal quantity to order, safety inventory, reorder point and total inventory management costs. In addition, information was shown about the company where the improvements were implemented. In the second chapter, topics were explained in relation to the research plan, making known the problem, the formulation of the hypothesis, the general objective and the specific ones. In the third chapter, the methodology of the research was discussed, where the approach and design, the population, the definition of variables and the data collection were determined. In the fourth chapter, the research was developed calculating avoidable costs in 3 levels of service. Two scenarios were established (with avoidable costs and no avoidable costs) to the sub-variables service levels, optimal quantity to be ordered, security inventory, reorder point and total inventory management costs. Finally, the fifth chapter showed the analysis of results answering the problem if there is a correlation between the avoidable costs in maritime import operation and the level of service provision, and we will provide conclusions and recommendations on the research developed. / Tesis

Le développement et l'application de modèles de réapprovisionnement d’articles périssables est l'une des principales préoccupations des experts en la matière, le nombre et la variété des produits périssables augmentant de façon spectaculaire. L'une des lacunes majeures dans la littérature pour la gestion des produits périssables est que les chercheurs n'ont pas accordé suffisamment d'attention à deux aspects importants dans leurs modèles: i) les conditions stochastiques ; en particulier le délai stochastique est presque négligé car rendant les défis mathématiques plus compliqués ; ii) l'élaboration de politiques innovantes de réapprovisionnement prenant en compte les critères environnementaux ; en particulier la minimisation des émissions de CO2 comme second objectif dans un contexte de modélisation multi-objectif qui est tout à fait nouvelle. Dans cette thèse, nous étudions les politiques de réapprovisionnement pour les produits périssables sous conditions stochastiques sous forme de trois problématiques différentes. Dans la première, nous développons un modèle de réapprovisionnement à révision continue (r, Q) pour un détaillant qui offre un produit périssable en prenant en compte : un horizon de planification infini, un délai d’approvisionnement stochastique, un taux de demande constante et la livraison tardive (backorder). Pour modéliser le processus de détérioration, un coût de possession de stock non linéaire est défini. La prise en considération du délai stochastique et d'un coût de possession de stock non linéaire rend le modèle mathématique plus complexe. Nous avons donc adapté le modèle proposé pour une fonction de distribution uniforme afin de résoudre de façon optimale ce problème par une approche exacte. Pour le second problème, nous étudions la stratégie de mutualisation des risques de délai de livraison par la passation de commandes de réapprovisionnement fractionnées par lots entre plusieurs fournisseurs simultanément pour un détaillant vendant un produit périssable. Enfin, dans le dernier problème, nous prenons en considération les coûts de stockage et de transport, ainsi que les impacts sur l'environnement, dans une chaîne d'approvisionnement centralisée sous condition de demande incertaine et pénurie partielle (partial backordered). Pour faire face à l'incertitude de la demande, est adoptée une approche de programmation stochastique en deux étapes. Par la suite, en tenant compte de la capacité de transport de véhicules, nous développons un modèle mathématique de programmation mixte en nombres entiers. De cette façon, les meilleurs véhicules de transport et les politiques de réapprovisionnement sont déterminés par la recherche d'un équilibre entre les critères financiers et environnementaux. Un exemple numérique du monde réel est également présenté pour démontrer l'applicabilité et l'efficacité du modèle proposé. / The development and application of inventory models for deteriorating items is one of the main concerns of the experts in the domain, since the number and variety of deteriorating products are dramatically increasing. One of the major gaps in the deteriorating inventories literature is that researchers have not paid enough attention to two important features in their models: i) Considering stochastic conditions; especially stochastic lead time is almost overlooked since makes the mathematical challenges complicated, ii) designing innovative inventory policies by taking into account the environmental issues and particularly the CO2 emission as a new objective in a multi-objective framework that is quite new. In this thesis, we study replenishment policy for deteriorating products under stochastic conditions in form of three different problem areas. In the first one, we develop a continuous (r,Q) inventory model for a retailer that offers a deteriorating product by considering infinite planning horizon, stochastic lead time, constant demand rate and backordered shortages. For modeling the deterioration process, a non-linear holding cost is defined. Taking into consideration the stochastic lead time as well as a non-linear holding cost makes the mathematical model more complex. We therefore customize the proposed model for a uniform distribution function that could be tractable to solve optimally by an exact approach. In second problem, we study the strategy of pooling lead time risks by splitting replenishment orders among multiple suppliers simultaneously for a retailer that sells a deteriorating product. Finally, in the last problem, we consider inventory and transportation costs, as well as the environmental impacts in a centralized supply chain by taking into account uncertain demand and partial backordered shortages. In order to deal with demand uncertainty, a two stage stochastic programming approach is taken. Then, by considering transportation vehicles capacity, we develop a mixed integer mathematical model. In this way, the best transportation vehicles and replenishment policy are determined by finding a balance between financial and environmental criteria. A numerical example from the real world is also presented to show the applicability and effectiveness of the proposed model.

Diese Dissertation analysiert das optimale Zusammenspiel dynamischer Preissetzung, dynamischer Werbung und Bestandsmanagement. Wir betrachten verschiedene Optimierungsprobleme für einen monopolistischen Händler bei gegebener zeitabhängiger deterministischer Nachfrage. In Kapitel 2 erweitern wir das Modell von Rajan et al. (1992). Der Händler darf einen dynamischen Preis, eine dynamische Werberate und die Lagergröße bei fester Verkaufsdauer wählen, so dass der Barwert von Umsatz minus Lager-, Einkaufs- und (nichtlinearen) Werbekosten maximiert wird; zusätzlich zerfällt der Lagerbestand mit exponentieller Rate. Wir ermitteln die optimale Preis-Werbe-Steuerung und die optimale Lagergröße und betrachten auch semi-statische Situationen. Wir führen eine Sensitivitätsanalyse im Hinblick auf den Einfluss der Modellparameter auf die optimalen Ergebnisse durch und vergleichen die Ergebnisse des dynamischen Modells mit denen der semi-statischen Modelle. In Kapitel 3 interpretieren wir den Verkaufsprozess als gesteuerten Diffusionsprozess eines neuen Produktes und die Lagergröße als unerschlossenen Marktanteil. Der Anfangszustand ist exogen gegeben und die Nachfrage hängt zusätzlich vom gegenwärtigen Zustand des Systems ab. Ein Zerfall des Lagerbestandes und alle Kosten bis auf Werbekosten sind ausgenommen. Anders als in Helmes et al. (2013) leiten wir die optimale Steuerung mithilfe des Pontrjaginschen Maximumprinzips her. Als Anwendung betrachten wir das Modell von von Bertalanffy. In Kapitel 4 erweitern wir die Analyse von einperiodigen Modellen auf langfristige Modelle. Die Länge des Verkaufszyklus und die Lagergröße sind Entscheidungsvariablen, wobei die optimalen Steuerungen aus Kapitel 2 bzw. Kapitel 3 während eines Zyklus angewandt werden. Existenzbedingungen für ein optimales Paar aus Zykluslänge und Lagergröße werden hergeleitet. Wir analysieren verschiedene Anwendungs- und Illustrationsbeispiele und verifizieren Strukturaussagen der optimalen Entscheidungsgrößen. / This dissertation analyzes the optimal coordination of dynamic pricing, dynamic advertising, and inventory management. We consider different optimization problems for a monopolistic retailer who faces a time-dependent deterministic demand. In Chapter 2, we generalize the model of Rajan et al. (1992). The retailer is allowed to choose a dynamic price, a dynamic advertising rate, and the inventory capacity for a sales period of fixed length so that the present value of revenue minus inventory, purchasing and (nonlinear) advertising costs is maximized; in addition, the inventory deteriorates at an exponential rate. We derive the optimal dynamic price-advertising control and the optimal capacity and also consider partially static cases. For the optimally controlled dynamic model we carry out a sensitivity analysis with respect to the model parameters and we compare the results of the dynamic model with those of the partially static models. In Chapter 3, we interpret the sales process as the controlled adoption process of a new product and the inventory capacity as untapped market share. The initial state is assumed to be exogenously given and the demand depends on the current state of the system. We exclude, however, deterioration effects and any other costs but the cost of advertising. We derive the optimal controls using a different technique than Helmes et al. (2013) - we apply Pontryagin’s maximum principle. As an interesting application we consider the controlled von Bertalanffy model. In Chapter 4, we extend the analysis of one-period models to multi-period and longterm average models. Assuming that the optimal controls derived in Chapter 2 and Chapter 3 are applied throughout a cycle, we treat the cycle length and the capacity as decision variables. We derive conditions that ensure the existence of an optimal pair of cycle length and capacity. Various examples and illustrations are given, and structural properties of the optimal pair are verified.

This thesis focuses on reduction of costs associated with the acquisition of amorphous cores from China, namely reduction of transport costs. The first part introduces the company’s production program. Then the thesis provides theoretical knowledge and analysis of the current state of the ordering process. Based on the analysis I have proposed measures which are reducing transport costs. Conditions needed for the implementation of these proposals are proposed as well. The final part offers quantified benefits of each proposal.

Indian pharmaceutical companies spend one-third of their revenue from supply chain management (SCM) activities due to inherently poor transportation infrastructure. SCM is a vital function for many companies, as it is usually employed to lower expenses and increase sales for the company. SCM costs are higher in India than they are in other areas of the world, amounting to 13% of India's GDP. The purpose of this study was to explore SCM strategies Indian business leaders in the pharmaceutical industry have used to reduce the high costs associated with SCM. This study used a single case study research design and semistructured interviews to collect data from 3 SCM business leaders working in Indian pharmaceutical organizations and possessing successful experience in using SCM strategies to reduce high costs. Goldratt's (1990) theory of constraints was used as the conceptual framework for this study to identify challenges associated with SCM strategies. Data from semistructured interviews, observations, and company documents were processed and analyzed using data source triangulation, grouping the raw data into key themes. The following 3 themes emerged: distribution and logistics challenges, impact of SCM processes, and best practices and solutions. The implications for positive social change include the potential to reduce supply chain risk, which could lead to lower product prices for consumers, increased stakeholder satisfaction, and a higher standard of living.

Le but de cette thèse est de proposer des méthodes de gestion des approvisionnements adaptées à des contextes particuliers afin de minimiser les coûts logistiques engendrés dans un réseau logistique multi produits, multi niveaux confronté à une demande probabiliste. Au cours de cette thèse, nous nous sommes attachés à : - Proposer des méthodes de gestion des stocks et du transport pour des familles de produits dans différents contextes : - Une première politique de réapprovisionnement est proposée pour une famille de produits caractérisée par une demande aléatoire et répétitive. Cette politique est définie par un niveau de commande et par un niveau de ré-complètement de stock pour chaque produit et une période de réapprovisionnement. Dès qu’un produit atteint le niveau de commande, un réapprovisionnement de tous les produits de la famille est déclenché. - Une deuxième politique de réapprovisionnement est proposée pour une famille de produits caractérisée par une demande très aléatoire et ponctuelle. Cette politique est basée sur les ruptures de stock. A chaque rupture d’un produit présent dans le stock il y a déclenchement d’un réapprovisionnement de tous les produits de la famille. - Proposer une méthode de classification multicritères afin de constituer des groupes de produits relevant d’une politique donnée, chaque classe ou famille regroupant des produits réagissant identiquement. Cette classification des produits en familles homogènes permet d’identifier les caractéristiques déterminantes dans le choix des méthodes de gestion de stock et de transport. - Analyser et comparer les performances de ces deux politiques d’approvisionnement par rapport à des politiques de référence, ainsi que leur sensibilité au regard de quelques paramètres discriminants : variabilité de la demande ; coût des produits ; coût des commandes urgentes… / The aim of this work is to propose methods of inventory management tailored to specific contexts in order to minimize logistics costs incurred in a logistics network multi-product, multi-level faces a probabilistic demand. In my research work, we have focused on the following scientific key-issues: - The proposal of methods of inventory and transportation management for products families in different contexts. The first one is defined to solve the problems identified in a product family characterized by a variable and repetitive demand. This policy is defined by an ordering and a replenishment level for each product and a periodic review. As soon as a product reaches the ordering level, replenishment of the same family products is triggered. • The second one is dedicated to a product family which is characterized by a very punctual and variable demand. This policy is based on inventory shortages. Each shortage of a product in store triggers replenishment of all products of the same family. - The proposal of a multi-criteria classification method in order to select the suitable inventory management method according to specificities of the product and the demand - The analysis and performance comparison of the two procurement policies proposed with standard policies, and their sensitivity with respect to some parameters discriminating: demand variability, cost of products, cost of rush orders.

Αντικείμενο της διατριβής αποτελεί το πρόβλημα του προγραμματισμού παραγωγής και ελέγχου των αποθεμάτων των συστημάτων διακριτής παραγωγής. Στόχος της είναι να προτείνει ένα νέο σύστημα PPC (r-MRP) που θα ξεπερνά ορισμένους βασικούς περιορισμούς των MRP-based συστημάτων. Το σύστημα αυτό αναπτύχθηκε πάνω σε μία νέα μέθοδο που ενοποιεί τις λειτουργίες διαχείρισης αποθεμάτων και λεπτομερούς χρονοπρογραμματισμού σε μία ενιαία παράλληλη υπολογιστική διαδικασία (reverse MRP). Ο σχεδιασμός του προτεινόμενου συστήματος ξεπερνάει βασικές αδυναμίες των MRP-based συστημάτων, λαμβάνοντας υπόψη τους περιορισμούς δυναμικότητας των παραγωγικών πόρων, τη διακύμανση των χρόνων υστέρησης των εργασιών σε αυτούς και τις σχετικές προτεραιότητες των παραγγελιών, μέσα σε ένα περιβάλλον συνεχούς χρόνου. Επιπλέον εισάγει ένα εργαλείο αξιολόγησης και επιλογής ενός αποδοτικού συνδυασμού χρονοπρογράμματος εργοστασίου και πλάνου αποθεμάτων μέσα από ένα σύνολο διαθέσιμων και εφικτών εναλλακτικών σεναρίων. Η διαμόρφωση των εναλλακτικών αυτών σεναρίων γίνεται λαμβάνοντας κάθε φορά υπόψη τους υφιστάμενους περιορισμούς δυναμικότητας της επιχείρησης και των υλικών, καθώς και ένα σύνολο κριτηρίων, ενίοτε αλληλοσυγκρουόμενων. Τα εναλλακτικά σενάρια που σχηματίζονται αξιολογούνται μέσα από ένα σύνολο δεικτών απόδοσης τόσο χρονοπρογραμματισμού, όσο και διαχείρισης αποθεμάτων. Το προτεινόμενο σύστημα υλοποιήθηκε υπό τη μορφή πειραματικού λογισμικού, μέσα σε περιβάλλον προσομοίωσης διακριτών συμβάντων (discrete-event simulation). Η απόδοση του μελετήθηκε μέσα από την πιλοτική εφαρμογή του σε μία τυπική κλωστοϋφαντουργική μονάδα και βρέθηκε να υπερτερεί σε συγκεκριμένες περιπτώσεις της υπολογιστικής διαδικασίας ενός τυπικού συστήματος MRPII, όσον αφορά ένα σύνολο δεικτών απόδοσης χρονοπρογραμματισμού και διαχείρισης αποθεμάτων. Το προτεινόμενο σύστημα r-MRP θα μπορούσε να ενσωματωθεί σε ένα MRP-based σύστημα ως ένα διασυνδεόμενο υποσύστημα (module) υποστήριξης της λειτουργίας προγραμματισμού παραγωγής και ελέγχου αποθεμάτων. / This thesis discusses the design of a production planning and control system (PPC) for discrete manufacturing environments, referred to as r-MRP. The proposed approach operates under the framework of hierarchical finite capacity shop-floor modelling and discrete-event simulation. It overcomes some of the main deficiencies of MRP-based systems by considering capacity constraints, lead times fluctuations, lot sizing and priority control in a ‘bucket-less’ environment. The r-MRP system was developed based on a new method that integrates shop scheduling and material planning in a simultaneous computational process (reverse MRP). Under this methodology, material and capacity constraints are considered, together with a number of conflicting planning criteria, in a continuous time environment. Lead times are dynamically calculated variables, based on actual loads in each workcenter. A number of alternative feasible shop schedule and material plan combinations are formed in real time. The r-MRP system was implemented as an experimental software tool. Its performance has been evaluated in a case study of a textile company with ‘real life’ data. When studied through a set of simulation experiments it has been found to outperform standard MRPII system in cases of short production times, high value inventory items and frequent, small deliveries by suppliers, in terms of a number of scheduling and inventory performance indicators. The r-MRP tool can be incorporated in legacy IT systems as an interfaced functional module.

Le cadre général des travaux de cette thèse est la chaîne logistique des pièces de rechange et particulièrement l’amélioration de sa performance en prévision et en gestion des stocks à plusieurs niveaux de service. La spécificité d’une demande faible et erratique ne permet pas l’utilisation des approches classiques de prévision. Par ailleurs, les mesures de performance associées reposant sur des indicateurs purement statistiques ne reflètent pas nécessairement les objectifs de gestion d’une chaîne logistique qui cherche à s’aligner avec différents niveaux d’exigence des contrats de maintenance. Après une description des processus et des outils en prévision et en gestion de stocks utilisés dans ce contexte, ces travaux proposent de nouvelles approches de sélection des méthodes de prévisions qui intègrent prévision et gestion de stocks, qui permettent la priorisation en service ou en inventaire et qui utilisent une segmentation par profil de la demande et par niveau de maturité des pièces. Par la suite, ces travaux considèrent le cas de la différenciation client dans le modèle de gestion de stocks basé sur des prévisions, en comparant la performance de la politique à niveau critique avec les politiques classiques à stock commun ou séparé. Chaque processus étudié a été accompagné d’une analyse comparative entre différentes méthodes de prévisions et politiques de gestion de stocks sur des données réelles du cas d’application industrielle de la chaine logistique de GE Healthcare. Ce qui a permis de construire un ensemble de recommandations en fonction de la segmentation de la pièce et des priorités des décideurs. / The main scope of these works is the spare parts supply chain management, particularly the improvement of forecasting and inventory management performance. The specificity of low and erratic demand does not allow the use of conventional approaches of forecasting. Moreover, the associated performance measurements, based on purely statistical indicators, are not adapted to this context. Indeed, it should align with different levels of expectations of the maintenance contracts. After a presentation of the processes and tools used by the dedicated literature. Here we propose new performance analysis approaches seeking to combine the statistical performance of forecasting methods and inventory management performance while considering decision makers priorities and using demand pattern as well as parts age segmentations. These works also address the use of customer differentiation in the inventory model based on forecasting, by comparing the performance of the critical level policy with the common stock and separated stock policies. Each studied process was associated with a comparative analysis of different forecasting methods and inventory management models based on real data of the spare parts supply chain of GE Healthcare, which allowed us to define a set of methods and parameters recommendations according to both part segmentations and supply chain priorities.