Design choices and experience using real-world distributed systems

Aim:

This session is targeted to innovative solutions, applications, commercial systems and architectures for real-world distributed systems. Theoretical analyses as well as experimental studies are welcome.

Submission:

Authors are invited to submit through Easychair either

* full-length papers (maximum 15 pages) containing original results (to be published in Springer series "Studies in Computational Intelligence", as proceedings or post-proceedings)

or

* extended abstracts (maximum 4 pages) presenting work in progress or software prototypes (to be included in an electronic proceedings with ISBN).

All submissions must be in English and edited in Word, using the templates of the Springer book series

Topics:

Important Dates:

Program Committee:

Co-chairs:

Evolving from Natural Computing and Data Mining

Probabilistic Models and Metaheuristics for Scheduling

Scheduling and planning problems arises in many real-world contexts with various industrial and economic applications, from job shop scheduling and assembly line balancing to efficient cloud resource provisioning. Identifying good schedules usually leads to difficult optimization problems which involve constraints, multiple conflicting optimization criteria and uncertainty. The presence of uncertainty limits the effectiveness of deterministic models and methods. In the context of scheduling, the probabilistic models can be used as surrogates for evaluate the quality of a solution, can be used to estimate the likelihood of various events which influence the scheduling process or to incorporate prior knowledge about the problem into the search mechanism. On the other hand, the presence of constraints and of multiple optimization criteria limits the effectiveness of exact optimization methods, allowing the metaheuristics to prove their superiority for this class of problems.

The aim of this session is to group original contributions on scalable and robust scheduling metaheuristics based on appropriate probabilistic models and addressing various types of real world applications. Theoretical analyses as well as experimental studies are welcome.

The topics include but are not limited to:

• scheduling performance measures
• probabilistic models in scheduling (reliability/failure models, Bayesian networks, etc.)
• constructive scheduling heuristics versus iterative-improvement heuristics
• population-based scheduling metaheuristics (e.g. estimation of distribution algorithms, evolutionary algorithms, ant colony optimization, particle swarm optimization etc.)
• local search based scheduling metaheuristics (tabu search, variable neighborhood search, iterated local search, greedy randomized adaptive search, simulated annealing etc.)
• hybridization of metaheuristics with various local search heuristics
• scalability of scheduling algorithms
• real time scheduling
• energy aware scheduling
• applications in cloud computing, job-shop and flow-shop scheduling problems, vehicle routing problems, preventive maintenance scheduling etc.

Complex Networks and Landscape Analysis

Michael T. M. Emmerich (Leiden University) and Asep Maulana (Leiden University)

This track focuses on computational methods for complex network analysis and the closely related topic of landscape analysis.

In complex network analysis the subject of study are properties of networks the constituents of which are rather simple, but the dynamics are complex due to emergent behavior. Applications range from biological systems to the study of social networks and parallel processes in computer science. Typical problems are system identification and design, as well as finding explanations for emerging behavior and computational aspects.

The metaphor of a (fitness) landscape is used in biology and bio-inspired optimization to describe neighborhood systems equipped with one or many potential functions (e.g. fitness, free energy, objective function). The purpose in landscape analysis is to find stable points in biological and chemical systems, characterize problem difficulty in bio-inspired optimization, or to understand evolutionary dynamics. The canonical representation of landscapes in discrete configuration spaces are node-weighted networks, or generalizations of it (such as hypergraphs).

Both topics are closely related: Whereas in the first is more concerned with dynamical aspects the second deals with static analysis of networks. In this track we look for contribution to the topic of complex network and landscape analysis, including:

1. Computational methods for network identification and analysis
2. Finding motifs and patterns in large (dynamical) networks
3. Identification of biological and gene regulatory networks
4. Neighborhood systems, pre-topologies, and generalizations of graphs
5. Fast enumeration of networks and trees; combinatorial state space size
6. Design and optimization of complex networks
7. Landscape analysis: e.g. adaptive walk statistics, barrier trees, local optima networks,
8. Community Detection and Graph Clustering
9. Growth of networks; scale free networks; small worlds;
10. Real-world applications, e.g. in logistics and bioinformatics

Submissions:

Please submit your contribution before February 1, 2015 (check for updates on website).

Full papers and extended abstracts are possible forms of contribution. Every accepted paper will be orally presented on the EVOLVE 2015 event in Iasi and full papers will be published in the Springer proceedings. Extended abstracts will be published in the short paper proceedings with ISBN/ISSN. All papers will be reviewed in a single-blind review process.  Submission EVOLVE2015, EASYCHAIR via the main system and indicate the track in the submission.

http://www.evolve-conference.org

For questions please send email to Dr. Michael Emmerich (This email address is being protected from spambots. You need JavaScript enabled to view it. ).