Constraint programming technology allows optimisation experts to solve a broad category of personnel rostering problems, such as nurse rostering, airline crew rostering or retail worker scheduling. However, for problem domain experts to use this technology, the optimisation system must bridge the gap for users to easily explore solutions and influence constraints. Working with our energy industry partner for several years, we identified rostering problems involving multi-skilled shift workers present on site for extended periods. Their existing workflow for handling rostering (crew allocation), and rerostering (dealing with inevitable employee absences) and for time-limited formation of dedicated maintenance crews is labour intensive and complex, requiring in-depth knowledge of personnel files and skill competencies. To address this, we propose an interactive decision support system for crew rostering and rerostering, currently being deployed by our industry partner, that provides interactive tools for domain experts to perform exploration, validation, and conflict recovery.

We study the problem of training a roster of engineers, who are scheduled to respond to service calls that require a set of skills, and where engineers and calls have different locations. Both training an engineer in a skill and sending an engineer to respond a non-local service call incur a cost. Alternatively, a local contractor can be hired. The problem consists in training engineers in skills so that the quality of service (i.e. response time) is maximised and costs are minimised. The problem is hard to solve in practice partly because (1) the value of training an engineer in one skill depends on other training decisions, (2) evaluating training decisions means evaluating the schedules that are now made possible by the new skills, and (3) these schedules must be computed over a long time horizon, otherwise training may not pay off. We show that a monolithic approach to this problem is not practical. Instead, we decompose it into three subproblems, modelled with MiniZinc. This allows us to pick the approach that works best for each subproblem (MIP or CP) and provide good solutions to the problem. Data is provided by a multinational company.

Decision systems for solving real-world combinatorial problems must be able to report infeasibility in such a way that users can understand the reasons behind it, and understand how to modify the problem to restore feasibility. Current methods mainly focus on reporting one or more subsets of the problem constraints that cause infeasibility. Methods that also show users how to restore feasibility tend to be less flexible and/or problem-dependent. We describe a problem-independent approach to feasibility restoration that combines existing techniques from the literature in novel ways to yield meaningful, useful, practical and flexible user support. We evaluate the resulting framework on two real-world applications.

In this paper we present a new 3D line alignment technique that can be used in limited resourced MAVs for performing loop closures as well as mutual localisation between MAVs. We identify pairs of 3D line matches from RGB-D frames and find the optimal transformation which aligns these two line sets onto each other in order to find the corresponding relative pose. The alignment of the two 3D line sets are achieved by converting each line match into two matching point pairs and then subjecting them to a least squares minimization process. As we only maintain line features extracted from key frames, our method does not require large memory, high processing power nor a high communication bandwidth between robots. We validate our 3D line alignment technique by performing loop closures and mutual localisation on real-world datasets.

We present a computationally efficient RGB-D based pose estimation solution for less computationally resourced MAVs, which are ideally suited as members in a swarm. Our approach applies the sufficient statistics derived for a least-squares problem to our problem context. RANSAC-based outlier detection in aligning corresponding feature points is a time consuming operation in visual pose estimation. The additive nature of the used sufficient statistics significantly reduces the computation time of the RANSAC procedure since the pose estimation in each test loop can be computed by reusing previously computed sufficient statistics. This eliminates the need for recomputing estimates from scratch each time. A simpler hypotheses testing method gave similar performance in terms of speed but less accurate than our proposed method. We further increase the efficiency by reducing the problem size to four dimensions using attitude data from an Attitude and Heading Reference System (AHRS). Using a real-world dataset, we show that our algorithm saves up to 94% of computation time for the RANSAC-based procedure in pose estimation while improving the accuracy.

The primary goal of this work is to provide an easy and systematic way of developing safe soft real-time systems. To achieve this goal, we propose a method of generating real-time programs from formally verified models written as systems of timed automata. The models are verified using UPPAAL model checker prior to be processed by our code generators. A characteristic of our code generator is that the generated code runs in a non-real-time environment, i.e., a runtime environment without inherent real-time schedulers. To realize this, the code generator weaves timing checking code fragments within the generated programs. The generated code explicitly checks the real-time clock of its runtime to obey the timing constraints specified in the model. In this paper, we describe how to generate Java/C programs from UPPAAL timed automata and show the benefits of our method using a robot controller case study.