Insight SFI Research Centre for Data Analytics
http://hdl.handle.net/10468/2481
Insight SFI Research Centre for Data Analytics2021-05-05T02:07:54ZClassifier-based constraint acquisition
http://hdl.handle.net/10468/11237
Classifier-based constraint acquisition
Prestwich, Steven D.; Freuder, Eugene C.; O'Sullivan, Barry; Browne, David
Modeling a combinatorial problem is a hard and error-prone task requiring significant expertise. Constraint acquisition methods attempt to automate this process by learning constraints from examples of solutions and (usually) non-solutions. Active methods query an oracle while passive methods do not. We propose a known but not widely-used application of machine learning to constraint acquisition: training a classifier to discriminate between solutions and non-solutions, then deriving a constraint model from the trained classifier. We discuss a wide range of possible new acquisition methods with useful properties inherited from classifiers. We also show the potential of this approach using a Naive Bayes classifier, obtaining a new passive acquisition algorithm that is considerably faster than existing methods, scalable to large constraint sets, and robust under errors.
2021-04-17T00:00:00ZLeprechauns on the chessboard
http://hdl.handle.net/10468/11079
Leprechauns on the chessboard
Escamocher, Guillaume; O'Sullivan, Barry
We introduce in this paper leprechauns, fairy chess pieces that can move either like the standard queen, or to any tile within k king moves. We then study the problem of placing n leprechauns on an n×n chessboard. When k=1, this is equivalent to the standard n-Queens Problem. We solve the problem for k=2, as well as for k>2 and n≤(k+1)2, giving in the process an upper bound on the lowest non-trivial value of n such that the (k,n)-Leprechauns Problem is satisfiable. Solving this problem for all k would be equivalent to solving the diverse n-Queens Problem, the variant of the n-Queens Problem where the distance between the two closest queens is maximized. While diversity has been a popular topic in other constraint problems, this is not the case for the n-Queens Problem, making our results the first major ones in the domain.
2021-02-05T00:00:00ZComputing optimal (R,s,S) policy parameters by a hybrid of branch-and-bound and stochastic dynamic programming
http://hdl.handle.net/10468/11078
Computing optimal (R,s,S) policy parameters by a hybrid of branch-and-bound and stochastic dynamic programming
Visentin, Andrea; Prestwich, Steven D.; Rossi, Roberto; Tarim, S. Armagan
A well-known control policy in stochastic inventory control is the policy, in which inventory is raised to an order-up-to-level S at a review instant R whenever it falls below reorder-level s. To date, little or no work has been devoted to developing approaches for computing policy parameters. In this work, we introduce a hybrid approach that exploits tree search to compute optimal replenishment cycles, and stochastic dynamic programming to compute levels for a given cycle. Up to 99.8% of the search tree is pruned by a branch-and-bound technique with bounds generated by dynamic programming. A numerical study shows that the method can solve instances of realistic size in a reasonable time.
2021-01-13T00:00:00ZTheoretical models for underwater RFID and the impact of water salinity on the design of wireless systems
http://hdl.handle.net/10468/11111
Theoretical models for underwater RFID and the impact of water salinity on the design of wireless systems
Peres, Caroline; Pigeon, Melusine; Rather, Nadeem; Gawade, Dinesh; Buckley, John; Jafarzadeh, Hamed; O'Flynn, Brendan
Underwater wireless communications present challenges due to the characteristics of water as a propagation channel medium. Regardless, wireless communications are needed for a range of systems that operate underwater. Commonly used technologies for these use cases (radio-frequency, acoustic and optical communications) are lacking, as they generally suffer from strong attenuation, multipath effects and propagation delays. In this context, we explore the theoretical models for Path Loss of Radio Frequency Identification (RFID) systems underwater in regards to the salinity of the water. We also discuss RFID systems feasibility in such applications as aquaculture and fish stock management. This paper aims to discuss the theoretical transmission models for RFID systems underwater, separating them into near-field systems – which use Magnetic Induction (MI) to communicate – and far-field systems – that transfer data via Radio Frequency (RF). We determine the path loss for each case, the effect of the salinity in the model for the path loss, and present preliminary measurements of magnetic field strength underwater for different salinity values.
2020-12-30T00:00:00Z