This dissertation develops mathematical models to efficiently manage the inland waterway port operations while minimizing the overall supply chain cost. In the first part, a capacitated, multi-commodity, multi-period mixed-integer linear programming model is proposed capturing diversified inland waterway transportation network related properties. We developed an accelerated Benders decomposition algorithm to solve this challenging NP-hard problem. The next study develops a two-stage stochastic mixed-integer nonlinear programming model to manage congestion in an inland waterway transportation network under stochastic commodity supply and water-level fluctuation scenarios. The model also jointly optimizes trip-wise towboat and barge assignment decisions and different supply chain decisions (e.g., inventory management, transportation decisions) in such a way that the overall system cost can be minimized. We develop a parallelized hybrid decomposition algorithm, combining Constraint Generation algorithm, Sample Average Approximation (SAA), and an enhanced variant of the L-shaped algorithm, to effectively solve our proposed optimization model in a timely fashion. While the first two parts develop models from the supply chain network design viewpoint, the next two parts propose mathematical models to emphasize the port and waterway transportation related operations. Two two-stage, stochastic, mixed-integer linear programming (MILP) models are proposed under stochastic commodity supply and water level fluctuations scenarios. The last one puts the specific focus in modeling perishable inventories. To solve the third model we propose to develop a highly customized parallelized hybrid decomposition algorithm that combines SAA with an enhanced Progressive Hedging and Nested Decomposition algorithm. Similarly, to solve the last mathematical model we propose a hybrid decomposition algorithm combining the enhanced Benders decomposition algorithm and SAA to solve the large size of test instances of this complex, NP-hard problem. Both proposed approaches are highly efficient in solving the real-life test instances of the model to desired quality within a reasonable time frame. All the four developed models are validated a real-life case study focusing on the inland waterway transportation network along the Mississippi River. A number of managerial insights are drawn for different key input parameters that impact port operations. These insights will essentially help decisions makers to effectively and efficiently manage an inland waterway-based transportation network.

Dagvatten transporteras via dagvattenbrunnar in i ledningsnät och bort från betongytor. Problem uppstår då dessa sätter igen vilket leder till att vägar och bostäder kan översvämmas. Detta medför ett behov att övervaka när dagvattenbrunnarna sätter igen. Sundsvalls kommun har tillsammans med Mittuniversitet använt sig av differentiella tryckgivare för att mäta vattennivåer i syfte att detektera när dagvattenbrunnarna sätter igen. Tryckgivaren fungera bra under sommartid men ger felaktiga utslag under vintern. Det har därmed föreslagits en kontakt fri lösningsmetod som omfattar ToF LiDAR sensorn VL53L1X. ToF LiDAR sensorer används för att bestämma avstånd till objekt, eftersom dessa inte vanligtvis appliceras inom vattennivåmätningar är detta en ny metod i det området. Syftet med arbetet var att utreda ifall ToF LiDAR sensorn VL53L1X kunde användas för att mäta vattennivåer samt avgöra om den kan tillämpas för övervakning av dagvatten under sommar- och vintertid. Övergripande mål var att utreda ifall sensorn kunde implementeras i en nod. Ett flertal förstudier gjordes för att utreda vilka faktorer som påverkade mätresultatet och för att karakterisera sensorns konfigurering för att anpassa sensorn för vattennivåmätningar i avsikt att effektivisera mätmetoden. Det framgick av förstudierna att faktorerna vattengrumlighet, solljus och avstånd till mätobjektet påverkade mätningarna så att mätresultatet försämrades. Det har konstaterats med arbetet som underlag att sensorn kan mäta vattennivåer och kan tillämpas för att övervaka dagvattennivåer. Mätresultat vid vattennivåmätningar kan åstadkommas med en mätnoggrannhet på ca 28 mm och ett mätfel på ca 46 mm. Mätresultatet gäller under omständigheterna att vattnet är rent, under påverkan av solljus samt att sensorns höjdposition är maximalt 90 cm. / Stormwater is transported through stormwater wells into a passage system and away from concrete surfaces. Problem emerges when these clog and causes flooding on roads and housings. These convey a necessity for monitoring the wells so that they do not clog. Sundsvall municipality has together with Mittuniversitet, used differential pressure sensors to measure water levels to detect when the stormwater wells clog. The differential pressure sensor operates well under summer season but during winter, it gives inadequate readings. Therefore, a contact free method has been suggested, that comprises the ToF LiDAR sensor VL53L1X. ToF LiDAR sensors are applied to determine distances, and are not usually used to measure water levels, therefore the suggested method is novice method for these types of applications. The purpose of the study was to investigate if the ToF LiDAR sensor VL53L1X could be used to measure water levels and to decide if it can be applied for monitoring of stormwater during summer- and winter season. The overall purpose was to investigate if the sensor could be implemented in a node. Various pre-studies were done to examine what factors could influence the measurement results and to characterize the sensor configuration to adapt the sensor for water level measurements to increase the effectiveness of the measuring method. The outcome of the pre-studies was that the factors water turbidity, sunlight, and distance to measuring object influenced the measurements so that the measuring results became inferior. It has been established with the studies as ground that the sensor can measure water levels and that it can be applied to monitor stormwater levels. The measuring result when applied to water level measuring could be obtained with an accuracy of approximately 28 mm and an error of approximately 46 mm. The measuring result apply during the conditions that the water is clean, with influence of sunlight and given that the sensor height position is maximum 90 cm.