This paper introduces a novel automated decision-making system for integrating autonomous Mobility-on-Demand (AMOD) services with conventional public transport systems, focusing on two main optimization tasks: vehicle order matching (VOM) and vehicle relocation (RE). In VOM, the system decides which active orders are serviced by AMOD vehicles, assigns idle vehicles to these orders and decides which direct or combined routes with existing public transport should be executed. RE focuses on moving idle vehicles to better locations to boost network efficiency and ensure vehicles are optimally positioned for present and upcoming needs. Implemented in a Reinforcement Learning framework, this paper compares Q-Learning (QL) and Deep Reinforcement Learning (DRL) approaches to enhance operational efficiency in urban transport. The evaluation, conducted with real-world data from New York City, demonstrates that Reinforcement Learning significantly outperforms automated non-learning approaches, highlighting its suitability for enhancing AMOD services and their integration with existing public transportation systems.

Forecast combination is an established technique to improve forecast
accuracy and enterprise planning, where a key research question is (still) how to
weight individual forecasts. One common but largely unsuccessful approach is to
learn weights that minimize the mean squared error (MSE) on known observations,
usually from (instable) sample covariance matrices of past errors. These weights
are then shrunk to mitigate over-fitting and avoid high errors when using the
weights in novel forecasts. This can be done by shrinking the sample covariance
matrix to a less flexible matrix, e.g. the unit diagonal matrix, where even formulas
for the shrinkage level minimizing the expected deviation between the shrunk and
the true covariance matrix exist. We provide analyses with synthetic error data
showing that such shrink-levels generally not lead to MSE-minimizing weights
and argue that adjusted shrinkage criteria or machine-learning-based shrinkage
tuning is adviced to successfully apply such approaches in forecast combination.

In the global pursuit of sustainability, the role of private sector capital
is pivotal, necessitating a shift in financial flows towards sustainable initiatives.
However, the proliferation of greenwashing undermines trust among investors and
consumers alike. The detection and mitigation of greenwashing demand robust
automated systems. Nonetheless, this endeavor is challenging in absence of clear
definitions, limited availability of high quality data, and the qualitative nature of
sustainability claims. In contrast, the field of fake news detection, facing similar
obstacles, has undergone substantial methodological evolution. Employing a sys-
tematic mapping approach, we categorize existing concepts from the greenwash-
ing and fake news domain, unveiling similarities and distinctions. Our findings
contribute to the development of effective greenwashing detection methods by
establishing the conceptual foundation to transfer methodological advancements
from fake news detection to the realm of greenwashing. Furthermore, we offer
promising research avenues in the domains of greenwashing, fake news and their
intersection.