Rosiane de Freitas

Applying MGAP Modeling to the Hard Real-time Task Allocation on Multiple Heterogeneous Processors Problem

The usage of heterogeneous multicore platforms is appealing for applications, e.g. hard real-time systems, due to the potential reduced energy consumption offered by such platforms. However, the power wall is still a barrier to improving the processor design process due to the power consumption of components. Hard real-time systems are part of life critical environments and reducing the energy consumption on such systems is an onerous and complex process. This paper reassesses the problem of finding assignments of hard real-time tasks among heterogeneous processors taking into account timing constraints and targeting low power consumption. We also propose models based on a well-established literature formulation of the Multilevel Generalized Assignment Problem (MGAP). We tackle the problem from the perspective of different integer programming mathematical formulations and their interplay on the search for optimal solutions. Experimentation shows that using strict schedulability tests as constraints of 0/1 integer linear programming results in faster solvers capable of finding optimum solutions with lower power consumption.

Towards optimal solutions for the low power hard real-time task allocation on multiple heterogeneous processors

Abstract The usage of heterogeneous multicore platforms is appealing for applications, e.g. hard real-time systems, due to the potential reduced energy consumption offered by such platforms. However, even in such platforms the power wall phenomena still imposes limits to performance. Hard real-time systems are part of life critical environments and reducing the energy consumption on such systems is an onerous and complex process. We tackle the problem from the perspective of different representative integer programming mathematical formulations and their interplay on the search for optimal solutions for Rate Monotonic (RM) and Earliest Deadline First (EDF) scheduling algorithms. The proposed models are based on a well-established formulation in the operational research literature, namely, the Multilevel Generalized Assignment Problem (MGAP). This paper, therefore, assesses the problem of finding optimal allocations and frequency assignments of hard real-time tasks among heterogeneous processors targeting low power consumption, but taking into account timing constraints. Computational experiments show that finding optimal solutions reduces the estimated energy consumption of the evaluated cases when compared to state-of-the-art algorithms.

Facet-inducing inequalities and a cut-and-branch for the bandwidth coloring polytope based on the orientation model

Abstract The bandwidth coloring problem (BCP) is a generalization of the well-known vertex coloring problem (VCP), asking colors to be assigned to vertices of a graph such that the absolute difference between the colors assigned to adjacent vertices is greater than or equal to a weight associated to the edge connecting them. In this work we present an integer programming formulation for BCP based on the orientation model for VCP. We present two families of valid inequalities for this formulation, show that they induce facets of the associated polytope, and report computational experience suggesting that these families are useful in practice.

Reaching Optimum Solutions for the Low Power Hard Real-Time Task Allocation on Multiple Heterogeneous Processors Problem

The usage of heterogeneous multicore platforms is appealing for applications, e.g. hard real-time systems, due to the potential reduced energy consumption offered by such platforms. However, the power wall is still a barrier to improving the processor design process due to the power consumption of components. Hard real-time systems are part of life critical environments and reducing the energy consumption on such systems is an onerous and complex process. This paper assesses the problem of finding optimum allocations and frequency assignments of hard real-time tasks among heterogeneous processors targeting low power consumption but taking into account timing constraints. We also propose models based on a well-established formulation in the operational research literature of the Multilevel Generalized Assignment Problem (MGAP). We tackle the problem from the perspective of different integer programming mathematical formulations and their interplay on the search for optimal solutions for RM and EDF. Computational experiments show that providing upper bounds determined by a meta-heuristic based on genetic algorithm reduces the time to finding optimal solution from hours to milliseconds, enabling us to still pursue optimum in larger instances.

A hybrid genetic algorithm with local search approach for E/T scheduling problems on identical parallel machines

This work considers scheduling problems on single and parallel machines with arbitrary processing times and independent jobs, to minimize the sum of earliness-tardiness penalties. A Genetic Algorithm with a smart local search approach is presented, a 2-opt neighborhood-based with GPI moves and tie-breaking criteria, in a single sequence representation for single and multi-machine instances. Computational experiments are performed on Tanakas instances for single machine, achieving all optimal solutions obtained by an IP exact method, for 40, 50, and 100 jobs. Moreover, our method is also suitable for dealing with multi-machine instances, achieving good solutions in a reasonable execution time, for 40, 50, and 100 jobs, with 2, 4, and 10 machines.

Watch or Immerse?: Redefining Your Role in Big Shows

Great events as concerts, music festivals and football matches bring together many people from different backgrounds and interests. In these events, people act, most of the time, as passive spectators, with little or none interaction with the artists in the event. Nowadays, there is some effort to engage people in those spaces, as well as using simple technology such as objects that can be raised on pre-defined moments or, as most recently, using LED technology with sensors. This work proposes an interaction technology for crowds in great events using mobile devices. This technology involves a dynamic and an app to redefine crowd role in those events, increasing their participation and engagement. In order to evaluate people immersion in great events, a case study has been carried out in a controlled environment mimicking a music concert. As a result, it was observed that the participants felt engaged, immersed and motivated to use the proposed technology in music concerts or even other great events and they are also willing to use their own devices as a way of an active participation.

Designing Distributed Real-Time Systems to Process Complex Control Workload in the Energy Industry

The energy industry demands computing system technologies with advanced state-of-the-art techniques to achieve reliability and safety for monitoring and properly dealing with several complex constraints. These computing systems also require delivering correct data at the right time imposing hard real-time constraints, because there are lots of situations where missing critical data may be catastrophic. The challenges faced by computer engineers in the energy industry also include designing distributed real-time systems to process such complex control workload. Besides, the computing system may also demand high energy consumption on its own. In this chapter, we demonstrate how to construct a mathematical formulation applicable for these computing systems and how to solve it to distribute the hard real-time workload of the process control systems considering technological constraints and optimizing for low power consumption of such computing systems. We present two computational techniques of resolution: an exact algorithm based on Branch-and-Cut and a meta-heuristic based on Genetic Algorithm. While the exact algorithm combines a branch-and-cut strategy with response time based schedulability analysis, the genetic algorithm still considers the response time schedulability analysis but follows an evolutionary solving strategy. Both computational techniques deliver solutions for heterogeneous computing systems with a control application, considering precedence, preemption, mutual exclusion, timing, temperature, and capacity constraints. In computational experiments, we present the usage of such techniques in a case study based on a control system for a power plant monitoring application.

The Distance Polytope for the Vertex Coloring Problem

In this work we consider the distance model for the classical vertex coloring problem, introduced by Delle Donne in 2009. This formulation involves decision variables representing the distance between the colors assigned to every pair of distinct vertices, thus not explicitly representing the colors assigned to each vertex. We show close relations between this formulation and the so-called orientation model for graph coloring. In particular, we prove that we can translate many facet-inducing inequalities for the orientation model polytope into facet-inducing inequalities for the distance model polytope, and viceversa.

Virtual structures and heterogeneous nodes in dependency graphs for detecting metamorphic malware

The traditional way to identify malicious programs is to compare the code body with a set of previously stored code patterns, also known as signatures, extracted from already identified malware code. To nullify this identification process, the malware developers can insert in their creations the ability to modify the malware code when the next contamination process takes place, using obfuscation techniques. One way to deal with this metamorphic malware behavior is the use of dependency graphs, generated by surveying dependency relationships among code elements, creating a model that is resilient to code mutations. Analog to the signature model, a matching procedure that compares these graphs with a reference graph database is used to identify a malware code. Since graph matching is a NP-hard problem, it is necessary to find ways to optimize this process, so this identification technique can be applied. Using dependency graphs extracted from binary code, we present an approach to reduce the size of the reference dependency graphs stored on the graph database, by introducing a node differentiation based on its features. This way, in conjunction with the insertion of virtual paths, it is possible to build a virtual clique used to identify and dispose of less relevant elements of the original graph. The use of dependency graph reduction also produces more stable results in the matching process. To validate these statements, we present a methodology for generating these graphs from binary programs and compare the results achieved with and without the proposed approach in the identification of the Evol and Polip metamorphic malware.

Sphere intersection algorithms for Molecular Distance Geometry Problem

The problem of estimating the full three-dimensional structure of a molecule, determining the position in space of all the atoms that compose it, is called Molecular Distance Geometry Problem (MDGP). To do this from an incomplete set of distances is NP-hard computational problem, where to get a feasible solution in a reasonable execution time presenting interesting mathematical and computational challenges. In this work, continuous and discrete mathematical approaches to solve MDGP is revised, based on the analysis of two types of calculating of sphere intersection: solving nonlinear systems from interatomic Euclidean distance equations, or solving internal coordinate systems using matrix multiplication techniques. We adapted the Branch-and-Prune (BP) method considering four spheres intersection. Computational experiments using instances from PDB benchmark are performed, determining the 3D structure based on our theoretical assumptions in a competitive computational processing time.