The hidden information and characteristics embedded in electrical load profiles are indispensable for the effective planning and operation of power grids. Load forecasting plays a vitally important role in many applications for the electric industry, e.g., energy generation and transactions, load shedding and restoration, as well as infrastructure expansions. Based on historical data, accurate load forecasts provide a good reference for the needed load demand, which can increase the efficiency and revenues of the electricity generation and distribution companies. In parallel with load forecasting, load monitoring can identify various types and statuses of loads by disaggregating the total power consumption into individual appliance levels. A scientific procedure of load monitoring facilitates the establishment of user-profiles, power usage habits, and peak load shifting. This is beneficial for both the end-users and utilities, improving the overall efficiency of the power network.

This research encapsulates three pivotal works focusing on enhancing the accuracy and efficiency of predicting and monitoring electrical loads. The first work introduces an attention-based neural load forecasting model that utilizes an encoder-decoder RNN and BiLSTM for dynamic feature selection and similar temporal information adaptively, showing superior performance in short-term load forecasting tasks. The second piece builds upon this, offering a unifying framework that integrates time-varying feature weighting, hierarchical temporal attention, and feature-reinforced error correction. This framework has shown outstanding efficacy in electric load forecasting on public datasets. The third work pivots to Non-Intrusive Load Monitoring (NILM). It introduces MATNilm, a multi-appliance-task framework, which efficiently disaggregates total power usage with limited labeled data. It employs a training-efficient sample augmentation (SA) scheme and a shared-hierarchical-split structure with a two-dimensional attention mechanism, enhancing the recognition of spatio-temporal correlations among all appliances. Collectively, these contributions underscore the integration of advanced deep learning techniques to improve the efficiency, reliability, and accuracy of load forecasting and monitoring in power systems.

Extensive numerical simulations reveal that our proposed load forecasting framework surpasses several existing forecasting methods. We emphasized the critical roles of both the feature weighting mechanism and the error correction module in securing this superior performance. In relation to the NILM task, even with just a single day's training data and limited appliance operation profiles, our SA algorithm demonstrates test performance comparable to models trained with a comprehensive dataset. In tandem with our proposed model structure, the simulation results highlight that our approach offers a notably enhanced performance against numerous baseline models, reducing relative errors by over 50% on average.

The Federal Aviation Administration uses Air Traffic Management Initiatives (TMIs) to mitigate the consequences of aviation system capacity shortfalls, for example by delaying aircraft on the ground at their origin airports. In order to make more efficient use of National Airspace (NAS) resources, reduce delay costs, and increase the flexibility of NAS users to meet their operational needs, tremendous efforts have been made to design TMIs in a manner to encourage cooperation between the FAA and airlines. Airlines are offered opportunities to make choices such as cancelling flights and increasing delays on some flights while decreasing delays on others. However, there has been little study of airlines' resulting behavior. In this dissertation, we analyze choices made by airlines in response to TMIs and attempt to infer from these key features of airlines' preference structures. Two econometric models are specified and estimated. The first model focuses on airlines' flight cancellation decisions, and the second model examines airline requests to simultaneously re-assign arrival slots and cancel flights using Slot Credit Substitution (SCS) messages.

The cancellation model captures how airlines value delays of the subject flight itself and potential delay savings of other flights in making a flight cancellation decision. Aircraft size, along with segment frequency and load factor, are all significant factors in cancellation decisions; larger, fuller, and less frequent flights are less likely to be cancelled. Somewhat surprisingly, a higher average fare is found to increase cancellation probability. Hub-bound flights are found more likely to be cancelled than spoke-bound flights. The model also confirms airlines' hedging behavior in response to TMIs by preferentially cancelling short-haul flights. In addition, a piece wise linear specification of the utility function confirms that the delay impact is non-linear. Individual airline model reveals some consistent behavior as well as some differences in how different factors enter into cancellation decisions.

The SCS model captures airlines' tradeoff behavior in dealing with flight cancellations and delays. It confirms that cancelling flights decreases airlines' utility while reducing delays increases the utility. Moreover, airlines are sensitive to the aircraft size and average fare of flights in performing these actions. In this model, however, average fare has the expected sign. The model estimates that airlines are willing to cancel a flight if the cancellation can reduce around 100 minutes of delays on their other flights that are in the ground delay program.