Nous développons un modèle basé agents pour reproduire deux anomalies fréquemment observées sur les marchés financiers : distribution leptokurtique des rendements et ampleur de la volatilité irrégulière mais persistante de ces mêmes rendements. Notre but est de montrer de façon probante que ces anomalies pourraient être attribuées à une formation mimétique des anticipations des intervenants sur les marchés. Nous nous éloignons des développements récents dans le domaine des modèles modèles basés agents en finance pour proposer un modèle très simple, estimé à partir des traits statistiques saillants de l’indice français journalier CAC 40. L’hypothèse d’anticipations mimétiques peut ainsi être testée : elle n’est pas rejetée dans notre modélisation. / We develop an agent-based model to replicate two frequently observed anomalies in the financial markets: the fat tails and the clustered volatility of the distribution of the returns. Our goal is to show conclusively that these anomalies could be attributed to a mimetic formation of the expectations of the stakeholders in the markets. We did not follow the rencent developpments in the field of the ACE model in the finance, but we propose a very simple model which is estimated from the stylized facts of the French daily index CAC 40. The hypothesis of mimetic anticipations can thus be tested: it is not rejected in our modeling.

Submitted by Nadir Basilio (nadirsb@uninove.br) on 2018-08-09T18:45:01Z No. of bitstreams: 1 Francisco Moraes Gomes.pdf: 3240990 bytes, checksum: 17659c72fed5b3faf44d712b20b24e39 (MD5) / Made available in DSpace on 2018-08-09T18:45:02Z (GMT). No. of bitstreams: 1 Francisco Moraes Gomes.pdf: 3240990 bytes, checksum: 17659c72fed5b3faf44d712b20b24e39 (MD5) Previous issue date: 2017-06-23 / The aim of this research is to study the potential association between the reduction in electricity consumption in Line 2 – Green of the São Paulo Subway and the change of the Train Control System from the "fixed block" type to the "moving block" type. This study is necessary because the transport sector is one of the biggest consumers of energy and the energy is one of the highest costs of the metro system, so that any reduction must be studied and understood. Most of the previous studies published on energy efficiency at subways and trains are simulation studies, there is a few literature based on empirical data. The statistical approach was chosen because there is no installed equipment to directly measure this influence, since there are only measures of consumption at the entrances of the System and there is no individual measurement of each train. Part of the data used was obtained via official communication channels of the company, another part was obtained directly from the operational reports and other data was extracted from the capture of the communications between the trains and centralized control system. Several variables that can influence the consumption of the line and also the train speed profiles in the periods just before and after the System change have been verified. The study of the relationships among these variables was done and the multivariate regression analysis was performed. The results of the regression generated an equation model with significance for the variables: type of system, proportion of useof the fleets used and the number of trips made. In the regression process, the stepwise method was used and some variables were withdrawn when verifying that they had multicollinearity. The data extracted from the communication protocol were not used in the regression, but they did allow verifying an evidence of improvement in the speed profiles and also a reduction in the number of stops that the train made in the route, indicating a smaller interference between the train movements with the new system. The calculated savings obtained by the new system was around 2.8%, rising to about 7.8% on average when the new fleet was included, which only uses the "moving block" system. / Esta pesquisa tem por objetivo estudar a potencial associação da ocorrência de queda no consumo de energia elétrica na Linha 2-Verde do Metrô de São Paulo com a troca do Sistema de Controle dos Trens do tipo “bloco fixo” para o tipo “bloco móvel”. Este estudo se mostra necessário, pois o setor de transporte é um dos maiores consumidores de energia e a energia é um dos maiores custos do sistema metroviário, de forma que qualquer redução deve ser estudada e compreendida. A maioria dos estudos anteriores publicados sobre eficiência energética em metrôs e trens são estudos de simulação, havendo pouca literatura baseada em dados empíricos. A abordagem estatística foi escolhida porque não há equipamentos instalados para se medir diretamente essa influência, pois apenas há medida de consumo somente na entrada de energia do Sistema, e não medição individual de cada trem. Parte dos dados utilizados foi obtida via canais oficias de comunicação da companhia, outra parte foi obtida diretamente dos relatórios operacionais e outros dados foram extraídos da captura das comunicações entre os trens e sistema de controle centralizado. Foram verificadas diversas variáveis que podem influir no consumo da linha e também os perfis de velocidades dos trens nos períodos imediatamente anteriores e posteriores a troca do Sistema. Foi realizado o estudo das relações entre essas variáveis e feita análise por regressão multivariada. Os resultados da regressão geraram um modelo de equação com significância para as variáveis: tipo de sistema, proporção de uso uma das frotas e a quantidade de viagens realizadas. No processo de regressão foi usado o método stepwise e algumas variáveis foram retiradas ao verificar que apresentavam multicolinearidade. Os dados extraídos do protocolo de comunicação não foram usados na regressão, mas por meio deles foi possível verificar indícios de melhoria nos perfis de velocidade e também uma redução na quantidade de paradas que o trem realizava no percurso, indicando uma menor interferência entre as movimentações dos trens com o novo sistema. A economia calculada obtida pelo novo sistema ficou em torno de 2,8%, subindo para cerca de 7,8% em média ao se incluir a nova frota, a qual só utiliza o sistema de "bloco móvel".

Energy consumption reduction is one of the priorities of metro operators, due to financial cost and environmental impact. The new signalling system Communications-Based Train Control (CBTC) is being installed in new and upgraded metro lines to increase transportation capacity. But its continuous communication feature also permits to improve the energy performance of traffic operation, by updating the control command of the Automatic Train Operation (ATO) system at any point of the route. The present research addresses two main topics. The first is the design of efficient CBTC speed profiles for undisturbed train trajectory between two stations. The second takes into account the interaction between two consecutive trains under abnormal traffic conditions and proposes a tracking algorithm to save energy. In the first part of the research an off-line methodology to design optimal speed profiles for CBTC-ATO controlled trains is proposed. The methodology is based on a new multi-objective optimisation algorithm named NSGA-II-F, which is used to design speed profiles in such a way that can cover all the possible efficient solutions in a pseudo-Pareto front. The pseudo–Pareto front is built by using dominated solutions to make available a complete set of feasible situations in a driving scenario. The uncertainty in the passenger load is modelled as a fuzzy parameter. Each of the resulting speed profiles is obtained as a set of parameters that can be sent to the ATO equipment to perform the driving during the operation. The proposed optimisation algorithm makes use of detailed simulation of the train motion. Therefore, a simulator of the train motion has been developed, including detailed model of the specific ATO equipment, the ATP constraints, the traction equipment, the train dynamics and the track. A subsequent analysis considers the effect in the design of considering the regenerative energy flow between the train and the surrounding railway system. The second part of the research is focused on the proposal and validation of a fuzzy tracking algorithm for controlling the motion of two consecutive trains during disturbed conditions. A disturbed condition is understood as a change in the nominal driving command of a leading train and its consequences in the subsequent trains. When a train runs close enough to the preceding one, a tracking algorithm is triggered to control the distance between both trains. The following train receives the LMA (limit of movement authority) via radio, which is updated periodically as the preceding train runs. The aim of the proposed algorithm is to take actions in such a way that the following train could track the leading train meeting the safety requirements and applying an energy saving driving technique (coasting command). The uncertainty in the variations of the speed of the preceding train is modelled as a fuzzy quantity. The proposed algorithm is based on the application of coasting commands when possible, substituting traction/braking cycles by traction/coasting cycles, and hence saving energy. Both algorithms were tested and validated by using a detailed simulation program. The NSGA-II-F algorithm provided additional energy savings when compared to fixed block distance-to-go configurations, and giving a more even distribution of the solutions. The fuzzy tracking algorithm provides energy savings with a minor impact on running times while improving comfort, because of the reduction of the inefficient traction/braking cycles. / <p>QC 20170216</p>

Communications-based train control (CBTC) aims to improve headways in the railway systems while maintaining or even improving safety. By means of the CBTC systems, the exact position of a train is known more accurately than the traditional signalling systems. It is a continuous, automatic train control system utilizing high-resolution train location determination, independent from track circuits. The system makes use of the telecommunications between the train and track equipment for the traffic management and infrastructure control. Roslagsbanan is a narrow gauge urban railway system that connects Stockholm with the northern suburbs of the city and despite its age and limited capacity, it is one of the most used railways in Sweden. From the beginning, the entire track consisted of single-tracks where the trains had to wait for each other on special meeting tracks. Αn expansion plan is now underway and the authorities are building approximately 22 km of double-track where the trains will be able to meet. All work along the Roslagsbanan aims to increase capacity, safety and accessibility. With more double-tracks and an updated signal system, the risk of traffic disruptions will be reduced and it is possible to have more frequent traffic in the future. This master’s thesis investigates a CBTC implementation in Roslagsbanan according to the current expansion plan and comparisons are conducted between the current capacity of the ATC system. A simulation model is developed that calculates the train’s safe braking curve and safe braking distance and two case studies are examined. The first one: Calculation of the capacity of the train routes and examination a capacity increase by a minimum 20% during rush hour, taking into consideration the current expansion plan in double-tracks. The second one: Calculation of the capacity of the train routes considering double-tracks the entire route. In this way, differences could be seen between the traffic density with single and double-tracks. Several insights were gained during the realization of this project and as well as timetables have been created to visualize the findings of this thesis. The timetable’s regularity time is decreased to a significant degree and more frequent traffic can be accommodated. Furthermore, the capacity is definitely affected by the track/infrastructure design however, it is questionable whether this increase in capacity outweighs the economic factor of expanding the entire route to double-tracks. / Kommunikationsbaserad tågkontroll (CBTC) syftar till att förbättra framsteg i järnvägssystemen samtidigt som säkerheten bibehålls eller till och med förbättras. Med hjälp av CBTC-systemen är ett tågs exakta position känd mer exakt än de traditionella signalsystemen. Det är ett kontinuerligt, automatiskt tågkontrollsystem som använder högupplöst tåglägesbestämning, oberoende av spårledning. Systemet använder telekommunikation mellan tåg och spårutrustning för trafikledning och infrastrukturkontroll. Roslagsbanan är ett smalspårigt stadsjärnvägssystem som förbinder Stockholm med stadens norra förorter och trots sin ålder och begränsade kapacitet är det en av de mest använda järnvägarna i Sverige. Från början bestod hela spåret av enkelspår där tågen fick vänta på varandra på speciella mötesstationer. Ett utvecklingsprogram pågår nu och myndigheterna bygger cirka 22 km dubbelspår där tågen kommer att kunna mötas. Allt arbete längs Roslagsbanan syftar till att öka kapacitet, säkerhet och tillgänglighet. Med fler dubbelspår och ett uppdaterat signalsystem minskar risken för trafikstörningar och det är möjligt att ha mer frekvent trafik i framtiden. Detta examensarbete undersöker ett CBTC-genomförande i Roslagsbanan enligt den nuvarande expansionsplanen och jämförelser görs med ATC-systemets nuvarande kapacitet. En simuleringsmodell har utvecklats som beräknar tågets säkra bromskurva och säkra bromssträcka och två fallstudier undersöks. Den första: Beräkning av tåglinjernas kapacitet och undersökning av en kapacitetsökning med minst 20% under rusningstid med hänsyn till den nuvarande expansionsplanen i dubbelspår. Den andra: Beräkning av tåglinjernas kapacitet med tanke på dubbelspår hela rutten. På detta sätt kunde skillnader ses mellan trafiktätheten med enkel- och dubbelspår. Flera insikter gjordes under genomförandet av detta projekt och tidtabeller har skapats för att visualisera resultaten av denna avhandling. Tiden mellan tågen minskas i betydande grad och mer frekvent trafik kan tillgodoses. Dessutom påverkas kapaciteten definitivt av ban-/infrastrukturutformningen, men det är tveksamt om denna kapacitetsökning uppväger den ekonomiska faktorn att utöka hela rutten till dubbelspår.

本文首次提出應用新的模擬方法：定態(stationary) bootstrap來估計涉險值(Value-at-Risk, 以下簡稱VaR)。VaR是衡量投資組合市場風險(market risk)的工具，由於1990年代以來國際間對市場風險管理的重視，且VaR較傳統風險衡量指標容易瞭解，又考慮整個投資組合資產間相關性降低風險的效果，加上VaR可作為企業內控、主管機關監督、以及投資人評估企業營運狀況等指標，故已廣為實務界及學界所接受。目前幾種主要衡量VaR的方法，包括變異數—共變數法、歷史模擬法、蒙地卡羅模擬法、classical bootstrap法以及壓力測試法等，各有其假設限制及優缺點。其中，classical bootstrap在衡量VaR時，使用的假設比較少，似乎非常適合衡量VaR。但是classical bootstrap會割裂了資料自我相關的性質，較適用於獨立且相同分配的樣本。我們在本文中介紹修正classical bootstrap的方法：移動區塊(moving block) bootstrap以及定態bootstrap，並利用統計模擬的方式證明定態bootstrap適合用於時間序列資料，對於捕捉真實分配的能力很強。接著我們選取11檔上市公司股票建構投資組合，並利用classical bootstrap以及定態bootstrap來衡量1999年共266個交易日的VaR。實證結果支持定態bootstrap能夠正確地衡量VaR，且其結果與classical bootstrap有明顯的不同。定態bootstrap法是個比較合理的衡量VaR方法，因此，我們建議風險管理者可採用定態bootstrap 衡量VaR。