In some cases, heuristics may be transferred easily between different optimisation problems. This is the case if these problems are equivalent or dual (e.g., maximum clique and maximum independent set) or have similar objective functions. However, the link between problems can further be defined by the constraints that define them. This refining can be achieved by organising constraints into families and translating between them using gadgets. If two problems are in the same constraint family, the gadgets tell us how to map from one problem to another and which constraints are modified. This helps better understand a problem through its constraints and how best to use domain specific heuristics. In this position paper, we argue that this allows us to understand how to map between heuristics developed for one problem to heuristics for another problem, giving an example of how this might be achieved.

When should one pay the ferryman? When should one pay for delivery of a good if there are no institutions or these are too costly to enforce contracts? We suggest to break up the transaction into many small rounds of investment and payment. We show that the e?cient investment can be implemented in an e-subgame perfect equilibrium for any given e if there are su?ciently many rounds of investment. This shows that when the horizon is ?nite, the holdup problem that emerges from backwards induction is not robust. Equilibria with stable and robust strategies require more periods.

We report preliminary results of linking distinct parameter metrics and stability via a novel dynamical system stability characterization (BBIOS). We conduct EA trials to determine the extent of EA stability in parameter space neighborhoods defined by metrics. We capture EA performance loss due to perturbation.

Hyper-heuristics are an optimization methodology which 'search the space of heuristics' rather than directly searching the space of the underlying candidate-solution representation. Hyper-heuristic search has traditionally been divided into two layers: a lower problem-domain layer (where domain-specific heuristics are applied) and an upper hyper-heuristic layer, where heuristics are selected or generated. The interface between the two layers is commonly termed the "domain barrier". Historically this interface has been defined to be highly restrictive, in the belief that this is required for generality. We argue that this prevailing conception of domain barrier is so limiting as to defeat the original motivation for hyper-heuristics. We show how it is possible to make use of domain knowledge without loss of generality and describe generalized hyper-heuristics which can incorporate arbitrary domain knowledge.

In this paper, we propose a general agent-based distributed framework where each agent is implementing a different metaheuristic/local search combination. Moreover, an agent continuously adapts itself during the search process using a direct cooperation protocol based on reinforcement learning and pattern matching. Good patterns that make up improving solutions are identified and shared by the agents. This agent-based system aims to provide a modular flexible framework to deal with a variety of different problem domains. We have evaluated the performance of this approach using the proposed framework which embodies a set of well known metaheuristics with different configurations as agents on two problem domains, Permutation Flow-shop Scheduling and Capacitated Vehicle Routing. The results show the success of the approach yielding three new best known results of the Capacitated Vehicle Routing benchmarks tested, whilst the results for Permutation Flow-shop Scheduling are commensurate with the best known values for all the benchmarks tested. (C) 2016 The Authors. Published by Elsevier B.V.

This paper presents a Decision Support System (DSS) for maintenance cost optimisation at an Offshore Wind Farm (OWF). The DSS is designed for use by multiple stakeholders in the OWF sector with the overall goal of informing maintenance strategy and hence reducing overall lifecycle maintenance costs at the OWF. Two optimisation models underpin the DSS. The first is a deterministic model that is intended for use by stakeholders with access to accurate failure rate data. The second is a stochastic model that is intended for use by stakeholders who have less certainty about failure rates. Solutions of both models are presented using a UK OWF that is in construction as an example. Conclusions as to the value of failure rate data are drawn by comparing the results of the two models. Sensitivity analysis is undertaken with respect to the turbine failure rate frequency and number of turbines at the site, with near linear trends observed for both factors. Finally, overall conclusions are drawn in the context of maintenance planning in the OWF sector.

We describe the 'Blackboard' design pattern for metaheuristics which allows multiple agents to combine their expertise opportunistically to contribute towards a solution. Features of the Blackboard pattern may include heterogeneity of solution representations (e.g. both graph and permutation for the TSP) and asynchronous processing, the latter rendering the traditionally hard distinction between 'online' and 'offline' activity less significant.

We propose that the development of standardized, explicit, machine-readable descriptions of metaheuris- tics will greatly advance scientific progress in the field. In particular, we advocate a purely functional description of metaheuristics — separate from any metaphors that inspire them and with no hidden mechanisms. A recent policy statement in the Journal of Heuristics1 highlights the need for improved research practice for metaheuristics via increased transparency of implementation and understanding of the contribution of their component parts. We describe here how addressing these issues via explicit descriptions can also offer further benefits. Standardization and pure-functional descriptions promote a higher standard of rigor for both communication and reproducibility of results. The modularity of our proposed approach opens up opportunities to compose heuristics in novel ways, along with better support for parallel processing. Most significantly, it is the basis for a greater degree of mechanized reasoning: we discuss how this might support large-scale collaborative research activity, leading to automated dis- covery, mining and assembly of metaheuristics.

Jerry Swan, Steven Adriaensen, Mohamed Bishr, Edmund K. Burke, John A. Clark, Patrick De Causmaecker, Juanjo Durillo, Kevin Hammond, Emma Hart, Colin G. Johnson, Zoltan A. Kocsis, Ben Kovitz, Krzysztof Krawiec, Simon Martin, J. J. Merelo, Leandro L. Minku, Ender Özcan, Gisele L. Pappa, Erwin Pesch, Pablo Garcı́a-Sánchez, Andrea Schaerf, Kevin Sim, Jim E. Smith, Thomas Stützle, Stefan Voß, Stefan Wagner, Xin Yao.

technical-note Share on The 'composite' design pattern in metaheuristics Authors: John Woodward University of Stirling, Stirling, United Kingdom University of Stirling, Stirling, United KingdomView Profile , Jerry Swan University of Stirling, Stirling, United Kingdom University of Stirling, Stirling, United KingdomView Profile , Simon Martin University of Stirling, Stirling, United Kingdom University of Stirling, Stirling, United KingdomView Profile Authors Info & Claims GECCO Comp '14: Proceedings of the Companion Publication of the 2014 Annual Conference on Genetic and Evolutionary ComputationJuly 2014 Pages 1439–1444https://doi.org/10.1145/2598394.2609848Published:12 July 2014Publication History 8citation121DownloadsMetricsTotal Citations8Total Downloads121Last 12 Months9Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

There have been several papers published relating to the practice of benchmarking in machine learning and Genetic Programming (GP) in particular. In addition, GP has been accused of targeting over-simplified 'toy' problems that do not reflect the complexity of real-world applications that GP is ultimately intended. There are also theoretical results that relate the performance of an algorithm with a probability distribution over problem instances, and so the current debate concerning benchmarks spans from the theoretical to the empirical. The aim of this article is to consolidate an emerging theme arising from these papers and suggest that benchmarks should not be arbitrarily selected but should instead be drawn from an underlying probability distribution that reflects the problem instances which the algorithm is likely to be applied to in the real-world. These probability distributions are effectively dictated by the application domains themselves (essentially data-driven) and should thus re-engage the owners of the originating data. A consequence of properly-founded benchmarking leads to the suggestion of meta-learning as a methodology for automatically designing algorithms rather than manually designing algorithms. A secondary motive is to reduce the number of research papers that propose new algorithms but do not state in advance what their purpose is (i.e. in what context should they be applied). To put the current practice of GP benchmarking in a particular harsh light, one might ask what the performance of an algorithm on Koza's lawnmower problem (a favourite toy-problem of the GP community) has to say about its performance on a very real-world cancer data set: the two are completely unrelated.

The development of decision support systems acceptable for nurse rostering practitioners still presents a daunting challenge. Building on an existing nurse rostering problem, a set of fairness-based objective functions recently introduced in the literature has been extended. To this end, a generic agent-based cooperative search framework utilising new mechanisms is described, aiming to combine the strengths of multiple metaheuristics. These different metaheuristics represent individual planners' implicit procedures for improving rosters. The framework enables to explore different ways of assessing nurse rosters in terms of fairness objectives. Computational experiments have been conducted across a set of benchmark instances. The overall results indicate that the proposed cooperative search for fair nurse rosters outperforms each metaheuristic run individually.

Nurse rostering is a complex real world problem requiring the assignment of various shifts to the hospital personnel subject to a set of constraints. Fairness of work allocation is a major human resources issue which has not been sufficiently explored in this field. The quality of a satisfactory solution to a nurse rostering problem instance is usually evaluated based on quantitative criteria. Nevertheless, the individuals have different implicit preferences and availabilities. Hence, the perceived roster quality may not be fully aligned with these quantitative criteria. Therefore, it would be more appropriate to also consider qualitative measures for fair rosters. This study investigates new qualitative measures for fairness and investigates centralised and cooperative meta-heuristic approaches for optimising these measures on a set of existing benchmark instances.

Cooperative search provides a class of strategies to design more effective search methodologies through combining (meta-) heuristics for solving combinatorial optimisation problems. This area has been little explored in operational research. In this study, we propose a general agent-based distributed framework where each agent implements a (meta-) heuristic. An agent continuously adapts itself during the search process using a cooperation protocol based on reinforcement learning and pattern matching. Good patterns which make up improving solutions are identified and shared by the agents. This agentbased system aims to raise the level of generality by providing a flexible framework to deal with a variety of different problem domains. The proposed framework has been so far tested on Permutation Flow-shop Scheduling and Travelling Salesman Problem instances yielding promising results.

In an increasingly interconnected economy, small and medium-sized enterprises (SMEs) are required to think about managing their business processes with a Business Process Management System (BPMS) in order to align themselves with larger companies whose activities are already governed by a BPMS. Yet today it is difficult for them to integrate a BPMS adapted to their needs with the BPMS of their business partners. This paper proposes a solution to facilitate interoperability between BPMS whose targeted enterprises differ in size. As proof of our concepts, we set up a demonstrator that visually executes a logistics supply chain scenario. Finally, we establish a list of recommendations to avoid pitfalls.

Taking as starting points a number of Gumbrecht's ideas and assertions about the merits and values of sport, this article considers if and why it might be worthy of our attention; as a spectacle and focus of study. Gumbrecht's questions about sport and ancient identities, its irresistible ability to seize the imagination, capture and distract attention, and his intrigue at the power of the mass crowds and empty stadia are related to various examples of sport in Fascist Italy. The first dictatorship to fully appreciate its power, this article shows how sport was both actor and object, exactly why this made it so important to the regime, and just how, in fact, this supports many of Gumbrecht's points. This brief analysis of Fascist sport that was both vast, ruthlessly organised and very successful, considers international and domestic football, boxing and Italy's World Heavyweight Champion Primo Carnera, plus the importance of crowds, stadiums and sporting venues, in assessing the role of athletes and events in the creation of athletic drama that was represented as a national collective experience. In the process, it will demonstrate sport's role as an important historical tool, whether it and its athletes are praiseworthy or not.

A model of two coupled single-server queues is analyzed in the steady state. Each new arrival into queue 1 is assigned a random deadline. If a job has not started service before its deadline expires, it is transferred to queue 2. The problem is to determine the optimal value of the average deadline period, in order to minimize a given cost func- tion expressed in terms of average numbers of jobs present in the two queues. That problem can be solved numerically with arbitrary degree of accuracy. A simple heuristic which produces good estimates is also developed; its performance is compared with the optimal solution.