Valery A. Kostenko

A Parallel Algorithm of Simulated Annealing for Multiprocessor Scheduling

Algorithms of simulated annealing for solving problems of multiprocessor scheduling are considered, an approach to parallelization is proposed, and the results of comparisons between the classical sequential, sequential, and parallel algorithms of simulated annealing using a partition of the initial space of solutions into regions are presented.

Scheduling algorithms for real-time computing systems admitting simulation models

A major requirement imposed on the operation of a real-time computing (RTC) system is that the deadlines for the operation of application programs must be met. The violation of an operational deadline leads to a failure of an RTC system. In this context, the problem arises of ensuring the required accuracy of estimating the execution time of application programs. An approach is developed for the design of iterative scheduling algorithms in which the execution times of application programs are estimated using simulation models with a different degree of detail, which ensures the required accuracy of estimating the execution time of programs. The approach can be used to design iterative algorithms of the following classes: genetic, evolutionary, simulated annealing, random-search, and locally optimal algorithms.

Algorithm for synthesis of real-time systems under reliability constraints

The problem of constructing a real-time computing system that has a minimum number of processors is addressed. It is necessary that the system meets the deadlines of program execution and the system reliability requirements implying that the system must tolerate both hardware and software failures. The formal statement of this problem is presented, a method for its solution using an iterative scheduling algorithm based on the method of simulated annealing is proposed, and an experimental study of the proposed algorithm is conducted.

Algorithm for resource allocation in data centers with independent schedulers for different types of resources

An algorithm for assigning requests to physical recourses for data centers with independent schedulers for different types of resources (computational resources, network resources, and data storages) is considered. This algorithm is based on the combination of greedy and limited search strategies. The algorithm provides a required balance between the computational complexity and quality of assignments by limiting the maximum allowed search depth. Theoretical and experimental results of investigating its features are presented.

An algorithm for constructing single machine schedules based on ant colony approach

A mathematical statement of the problem of building a schedule of data exchange over a channel with centralized control is presented. This problem arises in the design of real-time management information systems (MISs). Various approaches to the design of ant colony algorithms are analyzed, and experimental comparison of the proposed algorithm with greedy algorithms is performed.

An Algorithm for Scheduling Exchanges over a Bus with Centralized Control and an Analysis of Its Efficiency

In certain real-time systems, in particular, in onboard systems, it is often required to schedule of exchanges over a centralized communication channel (bus). Usually, end devices for the bus are various processors, sensors, and monitoring devices. They exchange (usually periodically) measurement results and controlling information. These exchanges are made according to a static schedule, which cannot be modified dynamically in the course of the system operation. According to this schedule, channel programs are developed. When the schedule is made up, various constraints must be taken into account; they are determined by the admissible modes of operation of particular on-board system and by the type of the bus controller. Examples of such constraints are given by maximally acceptable length of a chain of jobs that continuously follow one after another and the minimal required time interval between the chains of jobs during which no exchanges are scheduled. A heuristic algorithm for scheduling exchanges is developed that can be adapted to a set of constraints determined by a particular on-board system. The principles underlying this algorithm were briefly formulated in [1]. It is based on the greedy principle of scheduling according to the local criteria of the earliest possible time of the job completion with the check of optimality according to this criterion at every step. If the result is not optimal, a heuristic procedure is invoked that chooses the next job to be put on the schedule. At each step (after the next job is put on the schedule), the closest clock tick of the bus is determined at which the execution of the next job can be started; at each step, we also correct the admissible time intervals for the execution of the jobs that are still unscheduled. To take into account the constraints determined by the operation mode of the on-board system and the type of the bus controller, the fulfillment of the constraints is checked at every step of the algorithm; the results of the checking are used to correct the nearest clock tick at which the next job can be started. In this paper, we also determine the conditions for the initial data under which the algorithm produces an optimal schedule. Numerical results are also discussed.

The Problem of Schedule Construction in the Joint Design of Hardware and Software

Data structures for direct and parametric representation of schedules are considered and corresponding systems of operations for correcting schedules are introduced. Those operations are used for designing iteration algorithms of schedule construction. It is proved that the direct and parametric representation of schedules and the corresponding correction operations allow the construction of iteration algorithms that ensure the transformation of an arbitrary schedule into an optimal one in a linear number of iteration steps relative to the number of scheduled tasks. Moreover, those algorithms ensure that schedules obtained at all iteration steps are feasible.

Resource allocation algorithm in data centers with a unified scheduler for different types of resources

An algorithm for mapping requests onto physical resources for data centers with a unified scheduler for different types of resources—computational, network, and data storage ones—is considered. The algorithm is based on a “compact” assignment of requested resources for the creation of a virtual network in the data center. The desired balance between the computational complexity and quality of the resultant maps can be specified by limiting the maximally acceptable search depth. Theoretical and experimental results of the investigation of this algorithm’s properties are presented.

Comparing various approaches to resource allocation in data centers

In this paper, abstractions for describing resource requests and physical resources of data centers are chosen. A mathematical model of a data center is developed; this model provides an opportunity for describing a wide class of data center architectures. In terms of this model, a mathematical formulation of the resource allocation problem is given that admits migration of virtual machines and replication of data storage elements. Resource allocation algorithms for data centers with a unified scheduler for all types of resources, algorithms for data centers with specific schedulers for each type of resources, and similar algorithms from the OpenStack platform are compared; the comparison results are presented.

Synthesizing Structures of Real-Time Computer Systems Using Genetic Algorithms

A genetic algorithm is suggested for synthesizing real-time computer systems. The emphasis is on adjusting the algorithm to specific features of the problem and justifying the decisions made. Results of the analysis of the algorithm are also presented. Parameters of the genetic algorithm that provide good solutions are chosen by using a computational experiment.