Linear programming for large-scale Markov decision problems

We consider the problem of controlling a Markov decision process (MDP) with a large state space, so as to minimize average cost. Since it is intractable to compete with the optimal policy for large scale problems, we pursue the more modest goal of competing with a low-dimensional family of policies. We use the dual linear programming formulation of the MDP average cost problem, in which the variable is a stationary distribution over state-action pairs, and we consider a neighborhood of a low-dimensional subset of the set of stationary distributions (defined in terms of state-action features) as the comparison class. We propose a technique based on stochastic convex optimization and give bounds that show that the performance of our algorithm approaches the best achievable by any policy in the comparison class. Most importantly, this result depends on the size of the comparison class, but not on the size of the state space. Preliminary experiments show the effectiveness of the proposed algorithm in a queuing application.

Mathematics, programming, and STEM

Learning mathematics is a complex and dynamic process. In this paper, the authors adopt a semiotic framework (Yeh & Nason, 2004) and highlight programming as one of the main aspects of the semiosis or meaning-making for the learning of mathematics. During a 10-week teaching experiment, mathematical meaning-making was enriched when primary students wrote Logo programs to create 3D virtual worlds. The analysis of results found deep learning in mathematics, as well as in technology and engineering areas. This prompted a rethinking about the nature of learning mathematics and a need to employ and examine a more holistic learning approach for the learning in science, technology, engineering, and mathematics (STEM) areas.

The logic of data-sense: Thinking through Learning Personalisation

Big Data and Learning Analytics’ promise to revolutionise educational institutions, endeavours, and actions through more and better data is now compelling. Multiple, and continually updating, data sets produce a new sense of ‘personalised learning’. A crucial attribute of the datafication, and subsequent profiling, of learner behaviour and engagement is the continual modification of the learning environment to induce greater levels of investment on the parts of each learner. The assumption is that more and better data, gathered faster and fed into ever-updating algorithms, provide more complete tools to understand, and therefore improve, learning experiences through adaptive personalisation. The argument in this paper is that Learning Personalisation names a new logistics of investment as the common ‘sense’ of the school, in which disciplinary education is ‘both disappearing and giving way to frightful continual training, to continual monitoring'.

Logic of aid in an intertemporal setting

The paper studies the welfare implications of temporary foreign aid in the context of a simple two-country model of trade. In addition to its usual effects, a transfer of income in one period is assumed to influence the preferences of the recipient country in the following period. The implied changes in the terms of trade over the two periods are consistent with a number of possible outcomes with respect to the intertemporal welfare of the donor, the recipient, and the world as a whole. Particular attention is devoted to the conditions for strict Pareto improvement and the circumstances under which temporary aid transactions are likely to occur.

Combining unsupervised and invigilated assessment of introductory programming

We compared student performance on large-scale take-home assignments and small-scale invigilated tests that require competency with exactly the same programming concepts. The purpose of the tests, which were carried out soon after the take home assignments were submitted, was to validate the students' assignments as individual work. We found widespread discrepancies between the marks achieved by students between the two types of tasks. Many students were able to achieve a much higher grade on the take-home assignments than the invigilated tests. We conclude that these paired assessments are an effective way to quickly identify students who are still struggling with programming concepts that we might otherwise assume they understand, given their ability to complete similar, yet more complicated, tasks in their own time. We classify these students as not yet being at the neo-Piagetian stage of concrete operational reasoning.

A new constraint programming approach for optimising a coal rail system

Because of the bottlenecking operations in a complex coal rail system, millions of dollars are costed by mining companies. To handle this issue, this paper investigates a real-world coal rail system and aims to optimise the coal railing operations under constraints of limited resources (e.g., limited number of locomotives and wagons). In the literature, most studies considered the train scheduling problem on a single-track railway network to be strongly NP-hard and thus developed metaheuristics as the main solution methods. In this paper, a new mathematical programming model is formulated and coded by optimization programming language based on a constraint programming (CP) approach. A new depth-first-search technique is developed and embedded inside the CP model to obtain the optimised coal railing timetable efficiently. Computational experiments demonstrate that high-quality solutions are obtainable in industry-scale applications. To provide insightful decisions, sensitivity analysis is conducted in terms of different scenarios and specific criteria. Keywords Train scheduling · Rail transportation · Coal mining · Constraint programming