As a tool serving other disciplines of enquiry, artificial intelligence (AI) offers the potential of a potent discovery, a design and analysis paradigm to address (new) questions in urban planning. This thematic issue raises a forum for cross-disciplinary dialogues at the intersection of urban planning and AI. Nine articles discuss both emerging use cases in urban planning practice and the relevant AI techniques being used and developed, as well as articulate the challenges associated. Future development of AI in urban planning shall address the ethical, inclusive, and just implications of AI applications for urban planning while navigating human and AI agents’ interactions and intra-actions to facilitate a better understanding of the intentions of AI development and use, and the impacts on the behaviour of designers and users in complex urban planning practices.@en

Optimization models used to make discrete decisions often contain uncertain parameters that are context-dependent and estimated through prediction. To account for the quality of the decision made based on the prediction, decision-focused learning (end-to-end predict-then-optimize) aims at training the predictive model to minimize regret, i.e., the loss incurred by making a suboptimal decision. Despite the challenge of the gradient of this loss w.r.t. the predictive model parameters being zero almost everywhere for optimization problems with a linear objective, effective gradient-based learning approaches have been proposed to minimize the expected loss, using the empirical loss as a surrogate. However, empirical regret can be an ineffective surrogate because empirical optimal decisions can vary substantially from expected optimal decisions. To understand the impact of this deficiency, we evaluate the effect of aleatoric and epistemic uncertainty on the accuracy of empirical regret as a surrogate. Next, we propose three novel loss functions that approximate expected regret more robustly. Experimental results show that training two state-of-the-art decision-focused learning approaches using robust regret losses improves test–sample empirical regret in general while keeping computational time equivalent relative to the number of training epochs. @en

ReLU neural networks have been modelled as constraints in mixed integer linear programming (MILP), enabling surrogate-based optimisation in various domains and efficient solution of machine learning certification problems. However, previous works are mostly limited to MLPs. Graph neural networks (GNNs) can learn from non-euclidean data structures such as molecular structures efficiently and are thus highly relevant to computer-aided molecular design (CAMD). We propose a bilinear formulation for ReLU Graph Convolutional Neural Networks and a MILP formulation for ReLU GraphSAGE models. These formulations enable solving optimisation problems with trained GNNs embedded to global optimality. We apply our optimisation approach to an illustrative CAMD case study where the formulations of the trained GNNs are used to design molecules with optimal boiling points.@en

Delftse Foundations of Computation is a textbook for a one quarter introductory course in theoretical computer science. It includes topics from propositional and predicate logic, proof techniques, set theory and the theory of computation, along with practical applications to computer science. It has no prerequisites other than a general familiarity with computer programming.@en

The Dutch housing market comprises three sectors: social-rented, private-rented, and owner-occupied. The contemporary market is marked by a shortage of supply and a large subsidised social sector. Waiting lists for social housing are growing, whereas households with incomes above the limit do not or cannot leave the social sector. Government policy and market regulations change frequently, not least for political reasons. In view of commonly recognised problems in the housing market, this article considers the ‘internal demand’ of those households that are dissatisfied with their current residence. We examine the effects of regulatory policy by means of an exploratory agent-based simulation. The results provide perspectives on how internal demand is impacted by regulations in a housing market that is suffering from a shortage, and allow decision makers to weigh the pros and cons of policy measures.@en

Robust Optimal Control (ROC) with adjustable uncertainties has proven to be effective in addressing critical challenges within modern energy networks, especially the reserve and provision problem. However, prior research on ROC with adjustable uncertainties has predominantly focused on the scenario of uncertainties modeled as continuous variables. In this paper, we explore ROC with binary adjustable uncertainties, where the uncertainties are modeled by binary decision variables, marking the first investigation of its kind. To tackle this new challenge, firstly we introduce a metric designed to quantitatively measure the extent of binary adjustable uncertainties. Then, to balance computational tractability and adaptability, we restrict control policies to be affine functions with respect to uncertainties, and propose a general design framework for ROC with binary adjustable uncertainties. To address the inherent computational demands of the original ROC problem, especially in large-scale applications, we employ strong duality (SD) and big-M-based reformulations to create a scalable and computationally efficient Mixed-Integer Linear Programming (MILP) formulation. Numerical simulations are conducted to showcase the performance of our proposed approach, demonstrating its applicability and effectiveness in handling binary adjustable uncertainties within the context of modern energy networks.

@en

Exploration in reinforcement learning remains a difficult challenge. In order to drive exploration, ensembles with randomized prior functions have recently been popularized to quantify uncertainty in the value model. There is no theoretical reason for these ensembles to resemble the actual posterior, however. In this work, we view training ensembles from the perspective of Sequential Monte Carlo, a Monte Carlo method that approximates a sequence of distributions with a set of particles. In particular, we propose an algorithm that exploits both the practical flexibility of ensembles and theory of the Bayesian paradigm. We incorporate this method into a standard Deep Q-learning agent (DQN) and experimentally show qualitatively good uncertainty quantification and improved exploration capabilities over a regular ensemble.@en

The way how the uncertainties are represented by sets plays a vital role in the performance of robust optimization (RO). This paper presents a novel approach leveraging machine learning (ML) techniques to construct data-driven uncertainty sets from historical uncertainty data for RO problems. The proposed method integrates Density-Based Spatial Clustering of Applications with Noise (DBSCAN), Gaussian Mixture Model (GMM), and Principle Component Analysis (PCA) systematically to eliminate the influence of uncertainty scenarios with low occurrence probability and generate a nonconvex uncertainty set that is a union of multiple basic subsets (box or ellipsoid) without sacrificing its computational tractability. In addition to presenting a comprehensive algorithm for uncertainty set development, this paper offers detailed guidelines for parameter tuning and performance analysis. By harnessing the well-established ML packages scikit-learn, a Python-based toolkit for implementing the proposed approach is also provided. Furthermore, a computationally efficient solution for a two-stage linear RO problem with the proposed data-driven uncertainty set is derived, alongside establishing a probabilistic guarantee of constraint satisfaction for out-of-sample uncertainties. Extensive numerical experiments, conducted on both synthetic and real-world datasets as well as an optimization-based control problem, are performed to demonstrate the efficacy of the proposed methodology.

@en

Surrogate modelling techniques such as Kriging are a popular means for cheaply emulating the response of expensive Computational Fluid Dynamics (CFD) simulations. These surrogate models are often used for exploring a parameterised design space and identifying optimal designs. Multi-fidelity Kriging extends the methodology to incorporate data of variable accuracy and costs to create a more effective surrogate. This work recognises that the grid convergence property of CFD solvers is currently an unused source of information and presents a novel method that, by leveraging the data structure implied by grid convergence, could further improve the performance of the surrogate model and the corresponding optimisation process. Grid convergence states that the simulation solution converges to the true simulation solution as the numerical grid is refined. The proposed method is tested with realistic multi-fidelity data acquired with CFD simulations. The performance of the surrogate model is comparable to an existing method, and likely more robust. More research is needed to explore the full potential of the proposed method. Code has been made available online at https://github.com/robertwenink/MFK-Extrapolation.

@en

Social commitments are recognized as an abstraction that enables flexible coordination between autonomous agents. We make these contributions. First, we introduce and formalize a concept of a maintenance commitment, a kind of social commitment characterized by a maintenance condition whose truthhood an agent commits to maintain. This concept of maintenance commitments enables us to capture a richer variety of real-world scenarios than possible using achievement commitments with a temporal condition. Second, we develop a rule-based operational semantics, by which we study the relationship between agents' achievement and maintenance goals, achievement commitments, and maintenance commitments. Third, we motivate a notion of coherence between an agents' achievement and maintenance cognitive and social constructs, and prove that, under specified conditions, the goals and commitments of both rational agents individually and of a multiagent system altogether are coherent. Fourth, we illustrate our approach with a detailed real-world scenario from an aerospace aftermarket domain.

@en

The adoption of new market mechanisms - vital to the better integration of flexible assets - depends on the fairness and nondiscrimination of the pricing rules. We consider a market setting with time-flexible unit energy buyers and sellers, that additionally submit their availability in time. The time-flexibility of the agents allows for different schedules to be equivalent with regard to social welfare, which can lead to arbitrary price differences, i.e. price discrimination. In this work, we demonstrate that non-discriminatory prices are not trivially defined in time-flexible settings, provide a definition of non-discrimination as consistent over equivalent outcomes, show that this concept does not conflict with individual rationality and, finally, compare our work to broader concepts of fairness from economic psychology.

@en

Recent work has shown potential in using Mixed Integer Programming (MIP) solvers to optimize certain aspects of neural networks (NNs). However the intriguing approach of training NNs with MIP solvers is under-explored. State-of-the-art-methods to train NNs are typically gradient-based and require significant data, computation on GPUs, and extensive hyper-parameter tuning. In contrast, training with MIP solvers does not require GPUs or heavy hyper-parameter tuning, but currently cannot handle anything but small amounts of data. This article builds on recent advances that train binarized NNs using MIP solvers. We go beyond current work by formulating new MIP models which improve training efficiency and which can train the important class of integer-valued neural networks (INNs). We provide two novel methods to further the potential significance of using MIP to train NNs. The first method optimizes the number of neurons in the NN while training. This reduces the need for deciding on network architecture before training. The second method addresses the amount of training data which MIP can feasibly handle: we provide a batch training method that dramatically increases the amount of data that MIP solvers can use to train. We thus provide a promising step towards using much more data than before when training NNs using MIP models. Experimental results on two real-world data-limited datasets demonstrate that our approach strongly outperforms the previous state of the art in training NN with MIP, in terms of accuracy, training time and amount of data. Our methodology is proficient at training NNs when minimal training data is available, and at training with minimal memory requirements—which is potentially valuable for deploying to low-memory devices.@en

Increasing global food demand, accompanied by the limited number of expert growers, brings the need for more sustainable and efficient horticulture. The controlled environment of greenhouses enable data collection and precise control. For optimally controlling the greenhouse climate, a grower not only looks at crop production, but rather aims at maximising the profit. However this is a complex, long term optimisation task. In this paper, Constraint Programming (CP) is applied to task of optimal greenhouse economic control, by leveraging a learned greenhouse climate model through a CP embedding. In collaboration with an industrial partner, we demonstrate how to model the greenhouse climate with an LSTM model, embed this LSTM into a CP optimisation framework, and optimise the expected profit of the grower. This data-to-decision pipeline is being integrated into a decision support system for multiple greenhouses in the Netherlands.

@en

In the tool coating field, scheduling of production lines requires solving an optimisation problem which we call the multi-choice two-dimensional shelf strip packing problem with time windows. A set of rectangular items needs to be packed in two stages: items are placed on shelves, which in turn are placed on one of several available strips. Crucially, the item's width depends on the chosen strip and each item is associated with a time window such that items can only be placed on the same shelf if their time windows overlap. In collaboration with an industrial partner, this real-world optimisation problem is tackled in this paper by both exact and heuristic methods. The exact method is an arc-flow-based integer linear programming formulation, solved with the commercial solver CPLEX. Experimental evaluation shows that this approach can solve instances to proven optimality with up to 20 different item sizes. Larger, more realistic instances are solved heuristically by an adaptive large neighbourhood search, using first fit and best fit decreasing approaches as repair heuristics. In this way, we obtain high-quality solutions with a remaining optimality gap below 3.3% for instances with up to 2000 different item sizes. The work reported is due to be incorporated into an end-to-end decision support system with the industrial partner.

@en

Current state-of-the-art airline planning models face computational limitations, restricting the operational applicability to problems of representative sizes. This is particularly the case when considering the uncertainty necessarily associated with the long-term plan of an aircraft fleet. Considering the growing interest in the application of machine learning techniques to operations research problems, this article investigates the applicability of these techniques for airline planning. Specifically, an Advantage Actor–Critic (A2C) reinforcement learning algorithm is developed for the airline fleet planning problem. The increased computational efficiency of using an A2C agent allows us to consider real-world-sized problems and account for highly-volatile uncertainty in demand and fuel price. The result is a multi-stage probabilistic fleet plan describing the evolution of the fleet according to a large set of future scenarios. The A2C algorithm is found to outperform a deterministic model and a deep Q-network algorithm. The relative performance of the A2C increases as more complexity is added to the problem. Further, the A2C algorithm can compute a multi-stage fleet planning solution within a few seconds@en

Understanding preferences and behaviours in road freight transport is valuable for planning and analysis. This paper proposes a data-driven vehicle routing and scheduling approach for use as a descriptive tool to study road freight transport activities. The model developed seeks to capture planners’ or drivers’ preferences in order to reproduce observed road freight activities. The model is based on a parametrized time-dependent vehicle routing problem whose parameters can be estimated from a set of observed planned tours. We propose a Bayesian optimization technique for parameter estimation of the model. Empirical results show that the model can fit real-world data accurately and synthesize tour flows close to reality.

@en

COVID-19 significantly influenced travel behaviours and public attitudes towards public transport. Various studies have illustrated complicated factors related to long-term travel behaviour, indicating difficulty in understanding and predicting post-pandemic long-term travel behaviour via traditional methods. In these complex circumstances, it is valuable to take advantage of social media data to obtain real-time public opinions to understand dynamic travel behaviour changes from the passenger perspective. The present study provides a means - leveraging Twitter data and state-of-art Natural Language Processing (NLP) technologies - to interpret the underlying associations among public attitude, COVID-19 trends and public travel behaviour. Concretely, New York City has been selected due to its dependence on public transit for daily commuting. More than 500K tweets have been collected from January 2019 to June 2022. Automated text mining, topic modelling, and sentiment analysis have been implemented in these contexts to identify dynamic public reactions. A consistently negative attitude to public transit is detected and five main topics, including derivative topics from COVID-19, are discovered within the COVID-19 duration. Policy makers and transit managers can use these topics to take onboard the public's concerns. The paper thus exemplifies how social media data and NLP technologies can support policy-making progress and can benefit other tasks in the transportation domain.

@en

We solve a challenging scheduling problem with parallel batch processing and two-dimensional shelf strip packing constraints that arises in the tool coating field. Tools are assembled on so-called planetaries (batches) before they are loaded into coating machines to get coated. The assembling is not trivial and must fulfil specific constraints, which we refer to as shelf strip packing constraints. Further, each tool is associated with a starting time window s.t. tools can only be put on the same planetary if their time window overlap. The objective is to minimise the makespan and the number of required planetaries. Since the problem naturally decomposes into scheduling and packing parts, we tackle the problem with a two-phase logic-based Benders decomposition approach. The master problem assigns items to batches. The first phase solves as subproblem the packing problem by checking if the assignment is feasible, whereas the second phase solves the scheduling subproblem. The approach is compared with a monolithic mixed integer linear programming approach as well as a monolithic constraint programming approach. Experimental evaluation shows that our proposed approach outperforms the state-of-the-art benchmarks by solving more instances to optimality in a shorter time.

@en

Industrial and academic interest converge on scheduling flow shops with sequence- and time-dependent maintenance. We posit that anticipatory, integrated scheduling of operational and maintenance tasks leads to superior performance to purely 'wait-then-fix' handling of the maintenance tasks. Motivated by an industrial problem with (sequence dependent) setup times, maximum separation constraints, and a combination of sequence- and time- dependent maintenance tasks, this paper introduces an integer programming solution, a constraint programming solution and a heuristic solution based on list scheduling. The motivating use case provides a unique combination of concerns that is to the best of our knowledge, not yet studied in the literature. We build on existing work where we can by extending models for sequence-dependent maintenance scheduling to accommodate sequence- and time-dependent maintenance scheduling and also propose other new models. We show the relative performances of our methods through empirical evaluations and also show significant improvements - up to 25% reduction in makespan - when compared to a reactive scheduling approach that does not consider maintenance in its planning. Based on our evaluations on exact methods, constraint programming models scale better than mixed integer programming models for this problem.

@en

Since combinatorial scheduling problems are usually NP-hard, this paper investigates whether machine learning (ML) can accelerate exact solving of a problem instance. We adopt supervised learning on a corpus of problem instances, to acquire a function that predicts the optimal makespan for a given instance. The learned predictor is invariant to the instance size as it uses statistics of instance attributes. We provide this prediction to a solving algorithm in the form of bounds on the objective function. Specifically, this approach is applied to the well-studied Cyclic Hoist Scheduling Problem (CHSP). The goal for a CHSP instance is to find a feasible schedule for a hoist which moves objects between tanks with minimal cyclic period. Taking an existing Constraint Programming (CP) model for this problem, and an exact CP-SAT solver, we implement a Deep Neural Network, a Random Forest and a Gradient Boosting Tree in order to predict the optimal period p. Experimental results find that, first, ML models (in particular DNNs), can be good predictors of the optimal p; and, second, providing tight bounds for p around the predicted value to an exact solver significantly reduces the solving time without compromising the optimality of the solutions.@en