Continuity of care (COC) refers to the delivery of seamless services, continuous caring relationships, and information sharing across care providers. A disruption in COC—that is, care fragmentation (CF)—is an important cause of inefficiency in the Canadian healthcare system; such disruption leads to increased healthcare costs and reduced quality of care. Addressing this issue is particularly challenging among older adults, who often have medically complex needs; such patients can require many care transitions across multiple care settings. An effective strategy for COC improvements is to optimize discharge planning among older adults. However, this is hampered by the imperfect understanding of older patients’ needs, which are associated with their health complexity. Therefore, making early predictions about the patients’ health complexity and incorporating this information into discharge planning decisions can potentially improve COC. In this thesis, I develop data-driven predictive–prescriptive analytics frameworks that leverage machine learning (ML) approaches and a rich, massive set of longitudinal data collected over a decade. The first essay in this dissertation studies the early prediction of older patients’ complexity in hospital pathways using ML. It also examines whether we can conduct accurate prognostics with current information on patient complexity. The second study examines how two common measures of patient complexity—multimorbidity and frailty—concurrently affect post-discharge readmission and mortality among older patients. It also investigates the dependency of the outcomes on other essential socio-demographic factors. Finally, the third study examines the feasibility of predicting patients at risk of fragmented readmission—that is, readmission to a different hospital than the initial one. It uses this predictive information to derive optimal policies for preventing CF while addressing disparities in the decision-making process. The findings highlight the feasibility, utility, and performance of predicting patient complexity and important adverse outcomes, potentially undermining COC. This thesis shows that advanced knowledge and explicit utilization of this information could support decision-making and resource planning toward a targeted allocation at the system level; moreover, it informs actions that affect patient-centered care transition at the service level to optimize patient outcomes and facilitate upstream discharge processes, thereby improving COC. / Thesis / Doctor of Philosophy (PhD) / The aging population in Canada is growing significantly relative to the population as a whole, and several challenges are involved in providing aging people with proper healthcare services. One of these challenges is disruptions in continuity of care. Older adults are often medically complex or frail; they may have multiple diseases and require many care transitions across healthcare settings. Poor continuity of care among these patients leads to health deterioration during care trajectories, resulting in reduced quality of care and increased healthcare costs and inefficiencies. This thesis includes three essays that provide practical insights and solutions regarding the issue of continuity of care disruptions, spanning from predicting the issue to strategies to prevent it in a data-driven manner.

Long-term care facility network in Ontario, and in Canada as a whole, encounters critical issues regarding balancing demand with capacity. Even worse, it is faced with rising demand in the coming years. Moreover, there is an urgent need to provide long-term care for patients in their own language (particularly French). This study proposes a dynamic Mixed-Integer Linear Programming model based on the current standing of the long-term care system in Ontario, which simultaneously optimizes the time and location of constructing new long-term care facilities, adjusting the capacity (namely, human resources and beds) of each facility dynamically, and the assignment of patients to the facilities based on their demand region, gender, language, and age group over a finite time horizon. We apply the diversity-support constraints, based on patients’ gender and language, to save patients from loneliness and to comply with the Canadian values of providing care. Finally, we validate the model by performing a case study in Hamilton, Ontario. An extensive set of numerical analyses are explored to provide deeper insights into the whole issue. One set of such analysis is an extensive simulation study to examine the effect of distributional uncertainty in some of the input parameters on the optimal results, hence providing a much more realistic understanding of the optimization model. / Thesis / Master of Science (MSc)

More and more of our physical world today is being monitored and controlled by so-called cyber-physical systems (CPSs). These are compositions of networked autonomous cyber and physical agents such as sensors, actuators, computational elements, and humans in the loop. Today, CPSs are still relatively small-scale and very limited compared to CPSs to be witnessed in the future. Future CPSs are expected to be far more complex, large-scale, wide-spread, and mission-critical, and found in a variety of domains such as transportation, medicine, manufacturing, and energy, where they will bring many advantages such as the increased efficiency, sustainability, reliability, and security. To unleash their full potential, CPSs need to be equipped with, among other features, the support for automated planning and control, where computing agents collaboratively and continuously plan and control their actions in an intelligent and well-coordinated manner to secure and optimize a physical process, e.g., electricity flow in the power grid. In today’s CPSs, the control is typically automated, but the planning is solely performed by humans. Unfortunately, it is intractable and infeasible for humans to plan every action in a future CPS due to the complexity, scale, and volatility of a physical process. Due to these properties, the control and planning has to be continuous and automated in future CPSs. Humans may only analyse and tweak the system’s operation using the set of tools supporting prescriptive analytics that allows them (1) to make predictions, (2) to get the suggestions of the most prominent set of actions (decisions) to be taken, and (3) to analyse the implications as if such actions were taken. This thesis considers the planning and control in the context of a large-scale multi-agent CPS. Based on the smart-grid use-case, it presents a so-called PrescriptiveCPS – which is (the conceptual model of) a multi-agent, multi-role, and multi-level CPS automatically and continuously taking and realizing decisions in near real-time and providing (human) users prescriptive analytics tools to analyse and manage the performance of the underlying physical system (or process). Acknowledging the complexity of CPSs, this thesis provides contributions at the following three levels of scale: (1) the level of a (full) PrescriptiveCPS, (2) the level of a single PrescriptiveCPS agent, and (3) the level of a component of a CPS agent software system. At the CPS level, the contributions include the definition of PrescriptiveCPS, according to which it is the system of interacting physical and cyber (sub-)systems. Here, the cyber system consists of hierarchically organized inter-connected agents, collectively managing instances of so-called flexibility, decision, and prescription models, which are short-lived, focus on the future, and represent a capability, an (user’s) intention, and actions to change the behaviour (state) of a physical system, respectively. At the agent level, the contributions include the three-layer architecture of an agent software system, integrating the number of components specially designed or enhanced to support the functionality of PrescriptiveCPS. At the component level, the most of the thesis contribution is provided. The contributions include the description, design, and experimental evaluation of (1) a unified multi-dimensional schema for storing flexibility and prescription models (and related data), (2) techniques to incrementally aggregate flexibility model instances and disaggregate prescription model instances, (3) a database management system (DBMS) with built-in optimization problem solving capability allowing to formulate optimization problems using SQL-like queries and to solve them “inside a database”, (4) a real-time data management architecture for processing instances of flexibility and prescription models under (soft or hard) timing constraints, and (5) a graphical user interface (GUI) to visually analyse the flexibility and prescription model instances. Additionally, the thesis discusses and exemplifies (but provides no evaluations of) (1) domain-specific and in-DBMS generic forecasting techniques allowing to forecast instances of flexibility models based on historical data, and (2) powerful ways to analyse past, current, and future based on so-called hypothetical what-if scenarios and flexibility and prescription model instances stored in a database. Most of the contributions at this level are based on the smart-grid use-case. In summary, the thesis provides (1) the model of a CPS with planning capabilities, (2) the design and experimental evaluation of prescriptive analytics techniques allowing to effectively forecast, aggregate, disaggregate, visualize, and analyse complex models of the physical world, and (3) the use-case from the energy domain, showing how the introduced concepts are applicable in the real world. We believe that all this contribution makes a significant step towards developing planning-capable CPSs in the future. / Mehr und mehr wird heute unsere physische Welt überwacht und durch sogenannte Cyber-Physical-Systems (CPS) geregelt. Dies sind Kombinationen von vernetzten autonomen cyber und physischen Agenten wie Sensoren, Aktoren, Rechenelementen und Menschen. Heute sind CPS noch relativ klein und im Vergleich zu CPS der Zukunft sehr begrenzt. Zukünftige CPS werden voraussichtlich weit komplexer, größer, weit verbreiteter und unternehmenskritischer sein sowie in einer Vielzahl von Bereichen wie Transport, Medizin, Fertigung und Energie – in denen sie viele Vorteile wie erhöhte Effizienz, Nachhaltigkeit, Zuverlässigkeit und Sicherheit bringen – anzutreffen sein. Um ihr volles Potenzial entfalten zu können, müssen CPS unter anderem mit der Unterstützung automatisierter Planungs- und Steuerungsfunktionalität ausgestattet sein, so dass Agents ihre Aktionen gemeinsam und kontinuierlich auf intelligente und gut koordinierte Weise planen und kontrollieren können, um einen physischen Prozess wie den Stromfluss im Stromnetz sicherzustellen und zu optimieren. Zwar sind in den heutigen CPS Steuerung und Kontrolle typischerweise automatisiert, aber die Planung wird weiterhin allein von Menschen durchgeführt. Leider ist diese Aufgabe nur schwer zu bewältigen, und es ist für den Menschen schlicht unmöglich, jede Aktion in einem zukünftigen CPS auf Basis der Komplexität, des Umfangs und der Volatilität eines physikalischen Prozesses zu planen. Aufgrund dieser Eigenschaften müssen Steuerung und Planung in CPS der Zukunft kontinuierlich und automatisiert ablaufen. Der Mensch soll sich dabei ganz auf die Analyse und Einflussnahme auf das System mit Hilfe einer Reihe von Werkzeugen konzentrieren können. Derartige Werkzeuge erlauben (1) Vorhersagen, (2) Vorschläge der wichtigsten auszuführenden Aktionen (Entscheidungen) und (3) die Analyse und potentiellen Auswirkungen der zu fällenden Entscheidungen. Diese Arbeit beschäftigt sich mit der Planung und Kontrolle im Rahmen großer Multi-Agent-CPS. Basierend auf dem Smart-Grid als Anwendungsfall wird ein sogenanntes PrescriptiveCPS vorgestellt, welches einem Multi-Agent-, Multi-Role- und Multi-Level-CPS bzw. dessen konzeptionellem Modell entspricht. Diese PrescriptiveCPS treffen und realisieren automatisch und kontinuierlich Entscheidungen in naher Echtzeit und stellen Benutzern (Menschen) Prescriptive-Analytics-Werkzeuge und Verwaltung der Leistung der zugrundeliegenden physischen Systeme bzw. Prozesse zur Verfügung. In Anbetracht der Komplexität von CPS leistet diese Arbeit Beiträge auf folgenden Ebenen: (1) Gesamtsystem eines PrescriptiveCPS, (2) PrescriptiveCPS-Agenten und (3) Komponenten eines CPS-Agent-Software-Systems. Auf CPS-Ebene umfassen die Beiträge die Definition von PrescriptiveCPS als ein System von wechselwirkenden physischen und cyber (Sub-)Systemen. Das Cyber-System besteht hierbei aus hierarchisch organisierten verbundenen Agenten, die zusammen Instanzen sogenannter Flexibility-, Decision- und Prescription-Models verwalten, welche von kurzer Dauer sind, sich auf die Zukunft konzentrieren und Fähigkeiten, Absichten (des Benutzers) und Aktionen darstellen, die das Verhalten des physischen Systems verändern. Auf Agenten-Ebene umfassen die Beiträge die Drei-Ebenen-Architektur eines Agentensoftwaresystems sowie die Integration von Komponenten, die insbesondere zur besseren Unterstützung der Funktionalität von PrescriptiveCPS entwickelt wurden. Der Schwerpunkt dieser Arbeit bilden die Beiträge auf der Komponenten-Ebene, diese umfassen Beschreibung, Design und experimentelle Evaluation (1) eines einheitlichen multidimensionalen Schemas für die Speicherung von Flexibility- and Prescription-Models (und verwandten Daten), (2) der Techniken zur inkrementellen Aggregation von Instanzen eines Flexibilitätsmodells und Disaggregation von Prescription-Models, (3) eines Datenbankmanagementsystem (DBMS) mit integrierter Optimierungskomponente, die es erlaubt, Optimierungsprobleme mit Hilfe von SQL-ähnlichen Anfragen zu formulieren und sie „in einer Datenbank zu lösen“, (4) einer Echtzeit-Datenmanagementarchitektur zur Verarbeitung von Instanzen der Flexibility- and Prescription-Models unter (weichen oder harten) Zeitvorgaben und (5) einer grafische Benutzeroberfläche (GUI) zur Visualisierung und Analyse von Instanzen der Flexibility- and Prescription-Models. Darüber hinaus diskutiert und veranschaulicht diese Arbeit beispielhaft ohne detaillierte Evaluation (1) anwendungsspezifische und im DBMS integrierte Vorhersageverfahren, die die Vorhersage von Instanzen der Flexibility- and Prescription-Models auf Basis historischer Daten ermöglichen, und (2) leistungsfähige Möglichkeiten zur Analyse von Vergangenheit, Gegenwart und Zukunft auf Basis sogenannter hypothetischer „What-if“-Szenarien und der in der Datenbank hinterlegten Instanzen der Flexibility- and Prescription-Models. Die meisten der Beiträge auf dieser Ebene basieren auf dem Smart-Grid-Anwendungsfall. Zusammenfassend befasst sich diese Arbeit mit (1) dem Modell eines CPS mit Planungsfunktionen, (2) dem Design und der experimentellen Evaluierung von Prescriptive-Analytics-Techniken, die eine effektive Vorhersage, Aggregation, Disaggregation, Visualisierung und Analyse komplexer Modelle der physischen Welt ermöglichen und (3) dem Anwendungsfall der Energiedomäne, der zeigt, wie die vorgestellten Konzepte in der Praxis Anwendung finden. Wir glauben, dass diese Beiträge einen wesentlichen Schritt in der zukünftigen Entwicklung planender CPS darstellen. / Mere og mere af vores fysiske verden bliver overvåget og kontrolleret af såkaldte cyber-fysiske systemer (CPSer). Disse er sammensætninger af netværksbaserede autonome IT (cyber) og fysiske (physical) agenter, såsom sensorer, aktuatorer, beregningsenheder, og mennesker. I dag er CPSer stadig forholdsvis små og meget begrænsede i forhold til de CPSer vi kan forvente i fremtiden. Fremtidige CPSer forventes at være langt mere komplekse, storstilede, udbredte, og missionskritiske, og vil kunne findes i en række områder såsom transport, medicin, produktion og energi, hvor de vil give mange fordele, såsom øget effektivitet, bæredygtighed, pålidelighed og sikkerhed. For at frigøre CPSernes fulde potentiale, skal de bl.a. udstyres med støtte til automatiseret planlægning og kontrol, hvor beregningsagenter i samspil og løbende planlægger og styrer deres handlinger på en intelligent og velkoordineret måde for at sikre og optimere en fysisk proces, såsom elforsyningen i elnettet. I nuværende CPSer er styringen typisk automatiseret, mens planlægningen udelukkende er foretaget af mennesker. Det er umuligt for mennesker at planlægge hver handling i et fremtidigt CPS på grund af kompleksiteten, skalaen, og omskifteligheden af en fysisk proces. På grund af disse egenskaber, skal kontrol og planlægning være kontinuerlig og automatiseret i fremtidens CPSer. Mennesker kan kun analysere og justere systemets drift ved hjælp af det sæt af værktøjer, der understøtter præskriptive analyser (prescriptive analytics), der giver dem mulighed for (1) at lave forudsigelser, (2) at få forslagene fra de mest fremtrædende sæt handlinger (beslutninger), der skal tages, og (3) at analysere konsekvenserne, hvis sådanne handlinger blev udført. Denne afhandling omhandler planlægning og kontrol i forbindelse med store multi-agent CPSer. Baseret på en smart-grid use case, præsenterer afhandlingen det såkaldte PrescriptiveCPS hvilket er (den konceptuelle model af) et multi-agent, multi-rolle, og multi-level CPS, der automatisk og kontinuerligt tager beslutninger i nær-realtid og leverer (menneskelige) brugere præskriptiveanalyseværktøjer til at analysere og håndtere det underliggende fysiske system (eller proces). I erkendelse af kompleksiteten af CPSer, giver denne afhandling bidrag til følgende tre niveauer: (1) niveauet for et (fuldt) PrescriptiveCPS, (2) niveauet for en enkelt PrescriptiveCPS agent, og (3) niveauet for en komponent af et CPS agent software system. På CPS-niveau, omfatter bidragene definitionen af PrescriptiveCPS, i henhold til hvilken det er det system med interagerende fysiske- og IT- (under-) systemer. Her består IT-systemet af hierarkisk organiserede forbundne agenter der sammen styrer instanser af såkaldte fleksibilitet (flexibility), beslutning (decision) og præskriptive (prescription) modeller, som henholdsvis er kortvarige, fokuserer på fremtiden, og repræsenterer en kapacitet, en (brugers) intention, og måder til at ændre adfærd (tilstand) af et fysisk system. På agentniveau omfatter bidragene en tre-lags arkitektur af et agent software system, der integrerer antallet af komponenter, der er specielt konstrueret eller udbygges til at understøtte funktionaliteten af PrescriptiveCPS. Komponentniveauet er hvor afhandlingen har sit hovedbidrag. Bidragene omfatter beskrivelse, design og eksperimentel evaluering af (1) et samlet multi- dimensionelt skema til at opbevare fleksibilitet og præskriptive modeller (og data), (2) teknikker til trinvis aggregering af fleksibilitet modelinstanser og disaggregering af præskriptive modelinstanser (3) et database management system (DBMS) med indbygget optimeringsproblemløsning (optimization problem solving) der gør det muligt at formulere optimeringsproblemer ved hjælp af SQL-lignende forespørgsler og at løse dem "inde i en database", (4) en realtids data management arkitektur til at behandle instanser af fleksibilitet og præskriptive modeller under (bløde eller hårde) tidsbegrænsninger, og (5) en grafisk brugergrænseflade (GUI) til visuelt at analysere fleksibilitet og præskriptive modelinstanser. Derudover diskuterer og eksemplificerer afhandlingen (men giver ingen evalueringer af) (1) domæne-specifikke og in-DBMS generiske prognosemetoder der gør det muligt at forudsige instanser af fleksibilitet modeller baseret på historiske data, og (2) kraftfulde måder at analysere tidligere-, nutids- og fremtidsbaserede såkaldte hypotetiske hvad-hvis scenarier og fleksibilitet og præskriptive modelinstanser gemt i en database. De fleste af bidragene på dette niveau er baseret på et smart-grid brugsscenarie. Sammenfattende giver afhandlingen (1) modellen for et CPS med planlægningsmulighed, (2) design og eksperimentel evaluering af præskriptive analyse teknikker der gør det muligt effektivt at forudsige, aggregere, disaggregere, visualisere og analysere komplekse modeller af den fysiske verden, og (3) brugsscenariet fra energiområdet, der viser, hvordan de indførte begreber kan anvendes i den virkelige verden. Vi mener, at dette bidrag udgør et betydeligt skridt i retning af at udvikle CPSer til planlægningsbrug i fremtiden.

More and more of our physical world today is being monitored and controlled by so-called cyber-physical systems (CPSs). These are compositions of networked autonomous cyber and physical agents such as sensors, actuators, computational elements, and humans in the loop. Today, CPSs are still relatively small-scale and very limited compared to CPSs to be witnessed in the future. Future CPSs are expected to be far more complex, large-scale, wide-spread, and mission-critical, and found in a variety of domains such as transportation, medicine, manufacturing, and energy, where they will bring many advantages such as the increased efficiency, sustainability, reliability, and security. To unleash their full potential, CPSs need to be equipped with, among other features, the support for automated planning and control, where computing agents collaboratively and continuously plan and control their actions in an intelligent and well-coordinated manner to secure and optimize a physical process, e.g., electricity flow in the power grid. In today’s CPSs, the control is typically automated, but the planning is solely performed by humans. Unfortunately, it is intractable and infeasible for humans to plan every action in a future CPS due to the complexity, scale, and volatility of a physical process. Due to these properties, the control and planning has to be continuous and automated in future CPSs. Humans may only analyse and tweak the system’s operation using the set of tools supporting prescriptive analytics that allows them (1) to make predictions, (2) to get the suggestions of the most prominent set of actions (decisions) to be taken, and (3) to analyse the implications as if such actions were taken. This thesis considers the planning and control in the context of a large-scale multi-agent CPS. Based on the smart-grid use-case, it presents a so-called PrescriptiveCPS – which is (the conceptual model of) a multi-agent, multi-role, and multi-level CPS automatically and continuously taking and realizing decisions in near real-time and providing (human) users prescriptive analytics tools to analyse and manage the performance of the underlying physical system (or process). Acknowledging the complexity of CPSs, this thesis provides contributions at the following three levels of scale: (1) the level of a (full) PrescriptiveCPS, (2) the level of a single PrescriptiveCPS agent, and (3) the level of a component of a CPS agent software system. At the CPS level, the contributions include the definition of PrescriptiveCPS, according to which it is the system of interacting physical and cyber (sub-)systems. Here, the cyber system consists of hierarchically organized inter-connected agents, collectively managing instances of so-called flexibility, decision, and prescription models, which are short-lived, focus on the future, and represent a capability, an (user’s) intention, and actions to change the behaviour (state) of a physical system, respectively. At the agent level, the contributions include the three-layer architecture of an agent software system, integrating the number of components specially designed or enhanced to support the functionality of PrescriptiveCPS. At the component level, the most of the thesis contribution is provided. The contributions include the description, design, and experimental evaluation of (1) a unified multi-dimensional schema for storing flexibility and prescription models (and related data), (2) techniques to incrementally aggregate flexibility model instances and disaggregate prescription model instances, (3) a database management system (DBMS) with built-in optimization problem solving capability allowing to formulate optimization problems using SQL-like queries and to solve them “inside a database”, (4) a real-time data management architecture for processing instances of flexibility and prescription models under (soft or hard) timing constraints, and (5) a graphical user interface (GUI) to visually analyse the flexibility and prescription model instances. Additionally, the thesis discusses and exemplifies (but provides no evaluations of) (1) domain-specific and in-DBMS generic forecasting techniques allowing to forecast instances of flexibility models based on historical data, and (2) powerful ways to analyse past, current, and future based on so-called hypothetical what-if scenarios and flexibility and prescription model instances stored in a database. Most of the contributions at this level are based on the smart-grid use-case. In summary, the thesis provides (1) the model of a CPS with planning capabilities, (2) the design and experimental evaluation of prescriptive analytics techniques allowing to effectively forecast, aggregate, disaggregate, visualize, and analyse complex models of the physical world, and (3) the use-case from the energy domain, showing how the introduced concepts are applicable in the real world. We believe that all this contribution makes a significant step towards developing planning-capable CPSs in the future. / Mehr und mehr wird heute unsere physische Welt überwacht und durch sogenannte Cyber-Physical-Systems (CPS) geregelt. Dies sind Kombinationen von vernetzten autonomen cyber und physischen Agenten wie Sensoren, Aktoren, Rechenelementen und Menschen. Heute sind CPS noch relativ klein und im Vergleich zu CPS der Zukunft sehr begrenzt. Zukünftige CPS werden voraussichtlich weit komplexer, größer, weit verbreiteter und unternehmenskritischer sein sowie in einer Vielzahl von Bereichen wie Transport, Medizin, Fertigung und Energie – in denen sie viele Vorteile wie erhöhte Effizienz, Nachhaltigkeit, Zuverlässigkeit und Sicherheit bringen – anzutreffen sein. Um ihr volles Potenzial entfalten zu können, müssen CPS unter anderem mit der Unterstützung automatisierter Planungs- und Steuerungsfunktionalität ausgestattet sein, so dass Agents ihre Aktionen gemeinsam und kontinuierlich auf intelligente und gut koordinierte Weise planen und kontrollieren können, um einen physischen Prozess wie den Stromfluss im Stromnetz sicherzustellen und zu optimieren. Zwar sind in den heutigen CPS Steuerung und Kontrolle typischerweise automatisiert, aber die Planung wird weiterhin allein von Menschen durchgeführt. Leider ist diese Aufgabe nur schwer zu bewältigen, und es ist für den Menschen schlicht unmöglich, jede Aktion in einem zukünftigen CPS auf Basis der Komplexität, des Umfangs und der Volatilität eines physikalischen Prozesses zu planen. Aufgrund dieser Eigenschaften müssen Steuerung und Planung in CPS der Zukunft kontinuierlich und automatisiert ablaufen. Der Mensch soll sich dabei ganz auf die Analyse und Einflussnahme auf das System mit Hilfe einer Reihe von Werkzeugen konzentrieren können. Derartige Werkzeuge erlauben (1) Vorhersagen, (2) Vorschläge der wichtigsten auszuführenden Aktionen (Entscheidungen) und (3) die Analyse und potentiellen Auswirkungen der zu fällenden Entscheidungen. Diese Arbeit beschäftigt sich mit der Planung und Kontrolle im Rahmen großer Multi-Agent-CPS. Basierend auf dem Smart-Grid als Anwendungsfall wird ein sogenanntes PrescriptiveCPS vorgestellt, welches einem Multi-Agent-, Multi-Role- und Multi-Level-CPS bzw. dessen konzeptionellem Modell entspricht. Diese PrescriptiveCPS treffen und realisieren automatisch und kontinuierlich Entscheidungen in naher Echtzeit und stellen Benutzern (Menschen) Prescriptive-Analytics-Werkzeuge und Verwaltung der Leistung der zugrundeliegenden physischen Systeme bzw. Prozesse zur Verfügung. In Anbetracht der Komplexität von CPS leistet diese Arbeit Beiträge auf folgenden Ebenen: (1) Gesamtsystem eines PrescriptiveCPS, (2) PrescriptiveCPS-Agenten und (3) Komponenten eines CPS-Agent-Software-Systems. Auf CPS-Ebene umfassen die Beiträge die Definition von PrescriptiveCPS als ein System von wechselwirkenden physischen und cyber (Sub-)Systemen. Das Cyber-System besteht hierbei aus hierarchisch organisierten verbundenen Agenten, die zusammen Instanzen sogenannter Flexibility-, Decision- und Prescription-Models verwalten, welche von kurzer Dauer sind, sich auf die Zukunft konzentrieren und Fähigkeiten, Absichten (des Benutzers) und Aktionen darstellen, die das Verhalten des physischen Systems verändern. Auf Agenten-Ebene umfassen die Beiträge die Drei-Ebenen-Architektur eines Agentensoftwaresystems sowie die Integration von Komponenten, die insbesondere zur besseren Unterstützung der Funktionalität von PrescriptiveCPS entwickelt wurden. Der Schwerpunkt dieser Arbeit bilden die Beiträge auf der Komponenten-Ebene, diese umfassen Beschreibung, Design und experimentelle Evaluation (1) eines einheitlichen multidimensionalen Schemas für die Speicherung von Flexibility- and Prescription-Models (und verwandten Daten), (2) der Techniken zur inkrementellen Aggregation von Instanzen eines Flexibilitätsmodells und Disaggregation von Prescription-Models, (3) eines Datenbankmanagementsystem (DBMS) mit integrierter Optimierungskomponente, die es erlaubt, Optimierungsprobleme mit Hilfe von SQL-ähnlichen Anfragen zu formulieren und sie „in einer Datenbank zu lösen“, (4) einer Echtzeit-Datenmanagementarchitektur zur Verarbeitung von Instanzen der Flexibility- and Prescription-Models unter (weichen oder harten) Zeitvorgaben und (5) einer grafische Benutzeroberfläche (GUI) zur Visualisierung und Analyse von Instanzen der Flexibility- and Prescription-Models. Darüber hinaus diskutiert und veranschaulicht diese Arbeit beispielhaft ohne detaillierte Evaluation (1) anwendungsspezifische und im DBMS integrierte Vorhersageverfahren, die die Vorhersage von Instanzen der Flexibility- and Prescription-Models auf Basis historischer Daten ermöglichen, und (2) leistungsfähige Möglichkeiten zur Analyse von Vergangenheit, Gegenwart und Zukunft auf Basis sogenannter hypothetischer „What-if“-Szenarien und der in der Datenbank hinterlegten Instanzen der Flexibility- and Prescription-Models. Die meisten der Beiträge auf dieser Ebene basieren auf dem Smart-Grid-Anwendungsfall. Zusammenfassend befasst sich diese Arbeit mit (1) dem Modell eines CPS mit Planungsfunktionen, (2) dem Design und der experimentellen Evaluierung von Prescriptive-Analytics-Techniken, die eine effektive Vorhersage, Aggregation, Disaggregation, Visualisierung und Analyse komplexer Modelle der physischen Welt ermöglichen und (3) dem Anwendungsfall der Energiedomäne, der zeigt, wie die vorgestellten Konzepte in der Praxis Anwendung finden. Wir glauben, dass diese Beiträge einen wesentlichen Schritt in der zukünftigen Entwicklung planender CPS darstellen. / Mere og mere af vores fysiske verden bliver overvåget og kontrolleret af såkaldte cyber-fysiske systemer (CPSer). Disse er sammensætninger af netværksbaserede autonome IT (cyber) og fysiske (physical) agenter, såsom sensorer, aktuatorer, beregningsenheder, og mennesker. I dag er CPSer stadig forholdsvis små og meget begrænsede i forhold til de CPSer vi kan forvente i fremtiden. Fremtidige CPSer forventes at være langt mere komplekse, storstilede, udbredte, og missionskritiske, og vil kunne findes i en række områder såsom transport, medicin, produktion og energi, hvor de vil give mange fordele, såsom øget effektivitet, bæredygtighed, pålidelighed og sikkerhed. For at frigøre CPSernes fulde potentiale, skal de bl.a. udstyres med støtte til automatiseret planlægning og kontrol, hvor beregningsagenter i samspil og løbende planlægger og styrer deres handlinger på en intelligent og velkoordineret måde for at sikre og optimere en fysisk proces, såsom elforsyningen i elnettet. I nuværende CPSer er styringen typisk automatiseret, mens planlægningen udelukkende er foretaget af mennesker. Det er umuligt for mennesker at planlægge hver handling i et fremtidigt CPS på grund af kompleksiteten, skalaen, og omskifteligheden af en fysisk proces. På grund af disse egenskaber, skal kontrol og planlægning være kontinuerlig og automatiseret i fremtidens CPSer. Mennesker kan kun analysere og justere systemets drift ved hjælp af det sæt af værktøjer, der understøtter præskriptive analyser (prescriptive analytics), der giver dem mulighed for (1) at lave forudsigelser, (2) at få forslagene fra de mest fremtrædende sæt handlinger (beslutninger), der skal tages, og (3) at analysere konsekvenserne, hvis sådanne handlinger blev udført. Denne afhandling omhandler planlægning og kontrol i forbindelse med store multi-agent CPSer. Baseret på en smart-grid use case, præsenterer afhandlingen det såkaldte PrescriptiveCPS hvilket er (den konceptuelle model af) et multi-agent, multi-rolle, og multi-level CPS, der automatisk og kontinuerligt tager beslutninger i nær-realtid og leverer (menneskelige) brugere præskriptiveanalyseværktøjer til at analysere og håndtere det underliggende fysiske system (eller proces). I erkendelse af kompleksiteten af CPSer, giver denne afhandling bidrag til følgende tre niveauer: (1) niveauet for et (fuldt) PrescriptiveCPS, (2) niveauet for en enkelt PrescriptiveCPS agent, og (3) niveauet for en komponent af et CPS agent software system. På CPS-niveau, omfatter bidragene definitionen af PrescriptiveCPS, i henhold til hvilken det er det system med interagerende fysiske- og IT- (under-) systemer. Her består IT-systemet af hierarkisk organiserede forbundne agenter der sammen styrer instanser af såkaldte fleksibilitet (flexibility), beslutning (decision) og præskriptive (prescription) modeller, som henholdsvis er kortvarige, fokuserer på fremtiden, og repræsenterer en kapacitet, en (brugers) intention, og måder til at ændre adfærd (tilstand) af et fysisk system. På agentniveau omfatter bidragene en tre-lags arkitektur af et agent software system, der integrerer antallet af komponenter, der er specielt konstrueret eller udbygges til at understøtte funktionaliteten af PrescriptiveCPS. Komponentniveauet er hvor afhandlingen har sit hovedbidrag. Bidragene omfatter beskrivelse, design og eksperimentel evaluering af (1) et samlet multi- dimensionelt skema til at opbevare fleksibilitet og præskriptive modeller (og data), (2) teknikker til trinvis aggregering af fleksibilitet modelinstanser og disaggregering af præskriptive modelinstanser (3) et database management system (DBMS) med indbygget optimeringsproblemløsning (optimization problem solving) der gør det muligt at formulere optimeringsproblemer ved hjælp af SQL-lignende forespørgsler og at løse dem "inde i en database", (4) en realtids data management arkitektur til at behandle instanser af fleksibilitet og præskriptive modeller under (bløde eller hårde) tidsbegrænsninger, og (5) en grafisk brugergrænseflade (GUI) til visuelt at analysere fleksibilitet og præskriptive modelinstanser. Derudover diskuterer og eksemplificerer afhandlingen (men giver ingen evalueringer af) (1) domæne-specifikke og in-DBMS generiske prognosemetoder der gør det muligt at forudsige instanser af fleksibilitet modeller baseret på historiske data, og (2) kraftfulde måder at analysere tidligere-, nutids- og fremtidsbaserede såkaldte hypotetiske hvad-hvis scenarier og fleksibilitet og præskriptive modelinstanser gemt i en database. De fleste af bidragene på dette niveau er baseret på et smart-grid brugsscenarie. Sammenfattende giver afhandlingen (1) modellen for et CPS med planlægningsmulighed, (2) design og eksperimentel evaluering af præskriptive analyse teknikker der gør det muligt effektivt at forudsige, aggregere, disaggregere, visualisere og analysere komplekse modeller af den fysiske verden, og (3) brugsscenariet fra energiområdet, der viser, hvordan de indførte begreber kan anvendes i den virkelige verden. Vi mener, at dette bidrag udgør et betydeligt skridt i retning af at udvikle CPSer til planlægningsbrug i fremtiden.

The effective implementation of infrastructure asset management systems within organizations that own, operate, and manage infrastructure assets is critical to address the main challenges facing the infrastructure industry (e.g., infrastructure ageing and deterioration, maintenance backlogs, strict regulatory operating conditions, limited financial resources, and losing valuable experience through retirements). Infrastructure asset management systems contain connectivity between major operational components and such connectivity can lead to systemic risks (i.e., dependence-induced failures). This thesis analyzes the asset management system as a network of connected components (i.e., nodes and links) to identify critical components exposed to systemic risks induced by information asymmetry and information overload. This thesis applies descriptive and prescriptive analytics strategies to address information asymmetry and information overload and predictive analytics is employed to enhance the resilience. Specifically, descriptive analytics was employed to visualize the key performance indicators of infrastructure assets ensuring that all asset management stakeholders make decisions using consistent information sources and that they are not overwhelmed by having access to the entire database. Predictive analytics is employed to classify the resilience key performance indicator pertaining to the forced outage rapidity of power infrastructure components enabling power infrastructure owners to estimate the rapidity of an outage soon after its occurrence, and thus allocating the appropriate resources to return the infrastructure to operation. Using predictive analytics allows decision-makers to use consistent and clear information to inform their decision to respond to forced outage occurrences. Finally, prescriptive analytics is applied to optimize the asset management system network by increasing the connectivity of the network and in turn decreasing the exposure of the asset management system to systemic risk from information asymmetry and information overload. By analyzing an asset management system as a network and applying descriptive-, predictive-, and prescriptive analytics strategies, this dissertation illustrates how systemic risk exposure, due to information asymmetry and information overload could be mitigated and how power infrastructure resilience could be enhanced in response to forced outage occurrences. / Thesis / Doctor of Science (PhD) / Effective infrastructure asset management systems are critical for organizations that own, manage, and operate infrastructure assets. Infrastructure asset management systems contain main components (e.g., engineering, project management, resourcing strategy) that are dependent on information and data. Inherent within this system is the potential for failures to cascade throughout the entire system instigated by such dependence. Within asset management, such cascading failures, known as systemic risks, are typically caused by stakeholders not using the same information for decision making or being overwhelmed by too much information. This thesis employs analytics strategies including: i) descriptive analytics to present only relevant and meaningful information necessary for respective stakeholders, ii) predictive analytics to forecast the resilience key performance indicator, rapidity, enabling all stakeholders to make future decisions using consistent projections, and iii) prescriptive analytics to optimize the asset management system by introducing additional information connections between main components. Such analytics strategies are shown to mitigate the systemic risks within the asset management system and enhance the resilience of infrastructure in response to an unplanned disruption.

The increasing demand for healthcare services, coupled with the challenges of managing budgets and navigating complex regulations, has underscored the need for sustainable and efficient healthcare delivery. In response to this pressing issue, this thesis aims to optimize hospital efficiency using Artificial Intelligence (AI) techniques. The focus extends beyond improving surgical intraoperative time to encompass preoperative and postoperative periods as well. The research presents a novel Prescriptive Analytics System (PAS) designed to enhance the Surgical Success Rate (SSR) in surgeries and specifically in high volume arthroplasty. The SSR is a critical metric that reflects the successful completion of 4-surgeries during an 8-hour timeframe. By leveraging AI, the developed PAS has the potential to significantly improve the SSR from its current rate of 39% at The Ottawa Hospital to a remarkable 100%. The research is structured around five peer-reviewed journal papers, each addressing a specific aspect of the optimization of surgical efficiency. The first paper employs descriptive analytics to examine the factors influencing delays and overtime pay during surgeries. By identifying and analyzing these factors, insights are gained into the underlying causes of surgery inefficiencies. The second paper proposes three frameworks aimed at improving Operating Room (OR) throughput. These frameworks provide structured guidelines and strategies to enhance the overall efficiency of surgeries, encompassing preoperative, intraoperative, and postoperative stages. By streamlining the workflow and minimizing bottlenecks, the proposed frameworks have the potential to significantly optimize surgical operations. The third paper outlines a set of actions required to transform a selected predictive system into a prescriptive one. By integrating AI algorithms with decision support mechanisms, the system can offer actionable recommendations to surgeons during surgeries. This transformative step holds tremendous potential in enhancing surgical outcomes while reducing time. The fourth paper introduces a benchmarking and monitoring system for the selected framework that predicts SSR. Leveraging historical data, this system utilizes supervised machine learning algorithms to forecast the likelihood of successful outcomes based on various surgical team and procedural parameters. By providing real-time monitoring and predictive insights, surgeons can proactively address potential risks and improve decision-making during surgeries. Lastly, an application paper demonstrates the practical implementation of the prescriptive analytics system. The case study highlights how the system optimizes the allocation of resources and enables the scheduling of additional surgeries on days with a high predicted SSR. By leveraging the system's capabilities, hospitals can maximize their surgical capacity and improve overall patient care.