基于乐观策略的并行离散事件模拟研究

并行离散事件模拟（PDES）又称分布式模拟,通过将一个离散事件模拟程序在多个处理器上并行执行来提高模拟性能。乐观策略在解决并行模拟中各模拟部分之间的同步关系时具有较好的性能。介绍了基于乐观策略的并行离散事件模拟的原理,讨论了存在的问题,并给出相应的解决方法。     

并行离散事件模拟的同步机制研究

逻辑模拟在设计新系统的过程中起着重要作用,通过计算机进行模拟可以实时反馈输出结果,及早发现潜在的问题,进而缩短设计周期,降低研发成本。并行离散事件模拟通过分散计算量到并行机或者网络的多个节点来减少模拟时间,被视为解决模拟速度问题的有效途径。在影响模拟性能的因素中,各并行子系统之间的同步问题是直接影响并行性能的关键因素之一。探讨了并行离散事件模拟的同步机制,介绍了其基本原理、特点及存在的问题,并阐述了可能的改进方法。     

A fixed point approach to parallel discrete event simulation

Discrete event simulation is viewed as solving a fixed point problem whose unknowns are infinite histories or streams of event and time information. Stream domains provide two notions of convergence, which correspond to the usual categorization of simulation methods. Metric convergence leads to optimistic parallel simulation (the classic event list mechanism turns out to be a specialization), and convergence in the sense of partial orders leads to conservative parallel simulation.     

Asynchronous parallel simulation of parallel programs

Parallel simulation of parallel programs for large datasets has been shown to offer significant reduction in the execution time of many discrete event models. The paper describes the design and implementation of MPI-SIM, a library for the execution driven parallel simulation of task and data parallel programs. MPI-SIM can be used to predict the performance of existing programs written using MPI for message passing, or written in UC, a data parallel language, compiled to use message passing. The simulation models can be executed sequentially or in parallel. Parallel execution of the models are synchronized using a set of asynchronous conservative protocols. The paper demonstrates how protocol performance is improved by the use of application-level, runtime analysis. The analysis targets the communication patterns of the application. We show the application-level analysis for message passing and data parallel languages. We present the validation and performance results for the simulator for a set of applications that include the NAS Parallel Benchmark suite. The application-level optimization described in the paper yielded significant performance improvements in the simulation of parallel programs, and in some cases completely eliminated the synchronizations in the parallel execution of the simulation model     

Parallel discrete event simulation with SIMULA

The area of Discrete Event Simulation (DES) is the least impacted by parallel processing even though most of its applications require tremendous amounts of processing time. The common approach of parallelizing individually special purpose programs leads to very limited improvements in performance. We propose here the parallelization of general DES applications written in SIMULA, as a part of an ongoing project that aims towards developing methodologies and architectures for parallel DES. SIMULA is a general purpose process oriented language whose structure allows the creation of processes which can participate in a quasi-parallel execution according to an interleaved fashion. Problems related to the process interference representation, parallel scheduling and process synchronization are defined and solved. To test and verify the theoretical results the parallel execution of experimental and real DES applications has been simulated. The results show that significant improvement in performance can be expected.     

Using Ada for discrete event simulation

The process interaction approach is proposed for developing a discrete simulation environment in Ada. The introduction of simulation facilities in Ada not only concerns the classical aspect of model building, but allows a new class of problems to be tackled, that is the testing of correctness of programs intended for real-time applications. In this paper attention is focused on the presentation of the process scheduling in the simulation context and on the definition of standard forms of interactions among processes. Simulation facilities are organized by making use of Ada's structuring concepts.     

Predicting the behavior of large scale P2P systems by parallel discrete event simulation

P2P systems are becoming the dominator of Internet.Such systems are typically composed of thousands to millions of physical computers,which make it difficult to predict their behaviors without a large scale distributed system simulator.This paper is an attempt to predict the behavior of large scale P2P systems by building a novel parallel simulator:AegeanSim,which provides parallel discrete event simulation of such systems on high performance server clusters.We abstract the execution of P2P applications wit...     

Quantization-based filtering in distributed discrete event simulation

Quantization is an approach to distributed logical simulation in which the value space is quantized and trajectories are represented by the crossings of a set of thresholds. This is an alternative to the common approach which discretizes the time base of a continuous trajectory to obtain a finite number of equally spaced sampled values over time. In distributed simulation, a quantizer checks for threshold crossings whenever an output event occurs and sends this value across to a receiver thereby reducing the number of messages exchanged among federates in a federation. This may increase performance in various ways such as decreasing overall execution time or allowing a larger number of entities to be simulated. Predictive quantization is a more advanced approach that sends just one bit of information instead of the actual real value size with the consequence that not only the number of messages, but also the message size, can be significantly reduced in this approach. In this paper, we present an approach to packaging individual bits into a large message packet, called multiplexed predictive quantization. We demonstrate that this approach can save significant overhead (thereby maximizing data transmission) and can reach close to 100% efficiency in the limit of large numbers of simultaneous message sources encapsulated within individual federates. We also discuss the tradeoff between message bandwidth utilization and the error incurred in the quantization. The results relate bandwidth utilization and error to quantum size for federations executing in the HLA-compliant discrete event distributed simulation environment, DEVS/HLA. The theoretical and empirical results indicate that quantization can be very scaleable due to reduced local computation demands as well as having extremely favorable network load reduction/simulation fidelity tradeoffs.     

Parallel discrete event simulation using shared memory

With traditional event-list techniques, evaluating a detailed discrete event simulation-model can often require hours or even days of computation time. By eliminating the event list and maintaining only sufficient synchronization to ensure causality, parallel simulation can potentially provide speedups that are linear in the numbers of processors. A set of shared-memory experiments using the Chandy-Misra distributed simulation algorithm, to simulate networks of queues is presented. Parameters of the study include queueing network topology and routing probabilities, number of processors, and assignment of network nodes to processors. These experiments show that Chandy-Misra distributed simulation is a questionable alternative to sequential simulation of most queuing network models.<>     

Web Services with generic simulation models for discrete event simulation

Today the Internet and the World Wide Web (WWW) are on the cusp of a paradigm shift. Up to now most actions in the WWW are sorts of human–computer interaction, but the introduction of the eXtensible Markup Language (XML) changed the perception. The Internet will be seen as a great space of information and with the use of XML and following technologies like Web Services, Grid Computing and Semantic Web the difference between human–machine interaction and machine–machine interaction vanishes. This work investigates the usefulness of XML in the simulation domain and uses Web Service technology to build the SimASP framework for discrete event simulation (DES).     

Predicting the performance of synchronous discrete event simulation

We develop a model to predict the performance of synchronous discrete event simulation. Our model considers the two most important factors for the performance of synchronous simulation: load balancing and communication. The effect of load balancing in a synchronous simulation is computed using probability distribution models. We derive a formula that computes the cost of synchronous simulation by combining a communication model called LogGP and computation granularity. Even though the formula is simple, it is effective in capturing the most important factors for the synchronous simulation. The formula helps us to predict the maximum speed up achievable by synchronous simulation. In order to examine the prediction model, we have simulated several large ISCAS logic circuits and a simple PCS network simulation on an SGI Origin 2000 and Terascale Computing System (TCS) at the Pittsburgh Supercomputing Center. The results of the experiment show that our performance model accurately predicts the performance of synchronous simulation. The performance model developed is used to analyze the effect of several factors that may improve the performance of synchronous simulation. The factors include problem size, load balancing, granularity, communication overhead, and partitioning.     

PL/I as a discrete event simulation tool

The programming language PL/I is considered as a most useful and portable discrete event simulation tool. This paper discusses design facilities required in simulation languages, making use of a model described in PL/I. The problem of queues, parallelism and synchronization in a simulator is solved. The structure of the simulator is considered.     

Predicting nuclear power-plant operator performance using discrete event simulation

A study was conducted to evaluate the use of discrete event simulation (DES) to predict human performance in a nuclear power plant control room environment. Computer simulation models of two disturbance scenarios were built using a simulation software program, Micro Saint. In parallel, data were also collected at a full-scope training simulator at the Halden man-machine laboratory (HAMMLAB) in Halden, Norway, using crews of commercial nuclear power plant operators from the Loviisa nuclear power plant in Loviisa, Finland. Comparisons were made between predicted operator performance data generated by the simulation models and crew performance in the HAMMLAB experiment to determine the degree of agreement between the simulated data and the data from operators. The models were then used to extrapolate advanced control room conditions and alarm systems that were not tested in the HAMMLAB experiment. This report summarizes these findings and provides recommendations for improvements to the DES approach for use by a regulatory agency.     

A context-based object-oriented application framework for discrete event simulation

Applying object-oriented programming (OOP) to construct simulation programs has gained momentum in the research community. Major research efforts involved with object-oriented paradigm adopted in discrete-event simulation are reviewed in this paper, along with a simulation application framework proposed which provides special reusability mechanisms. This object-oriented simulation framework consists of three object elements, that is, model, service and scheduler. The relations and interactions among these elements are also demonstrated. In addition, the abstraction of a model under such a framework is reached by a special concept and implementation, referred to as the simulation context. Each simulation context is used to model a component sub-system, e.g. the control or physical flows of a manufacturing system. A practical system, CSIMT+ +, is also developed to validate the feasibility of such a simulation application framework. Furthermore, various forms of application models based on this framework are also illustrated.     

Domain-specific discrete event modelling and simulation using graph transformation

Graph transformation is being increasingly used to express the semantics of domain-specific visual languages since its graphical nature makes rules intuitive. However, many application domains require an explicit handling of time to accurately represent the behaviour of a real system and to obtain useful simulation metrics to measure throughputs, utilization times and average delays. Inspired by the vast knowledge and experience accumulated by the discrete event simulation community, we propose a novel way of adding explicit time to graph transformation rules. In particular, we take the event scheduling discrete simulation world view and provide rules with the ability to schedule the occurrence of other rules in the future. Hence, our work combines standard, efficient techniques for discrete event simulation (based on the handling of a future event set) and the intuitive, visual nature of graph transformation. Moreover, we show how our formalism can be used to give semantics to other timed approaches and provide an implementation on top of the rewriting logic system Maude.     

Efficient autonomic cloud computing using online discrete event simulation

Interest is growing in open source tools that let organizations build IaaS clouds using their own internal infrastructures, alone or in conjunction with external ones. A key component in such private/hybrid clouds is virtual infrastructure management, i.e., the dynamic orchestration of virtual machines, based on the understanding and prediction of performance at scale, with uncertain workloads and frequent node failures. Part of the research community is trying to solve this and other IaaS problems looking to Autonomic Computing techniques, that can provide, for example, better management of energy consumption, quality of service (QoS), and unpredictable system behaviors. In this context, we first recall the main features of the NAM framework devoted to the design of distributed autonomic systems. Then we illustrate the organization and policies of a NAM-based Workload Manager, focusing on one of its components, the Capacity Planner. We show that, when it is not possible to obtain optimal energy-aware plans analytically, sub-optimal plans can be autonomically obtained using online discrete event simulation. Specifically, the proposed approach allows to cope with a broader range of working conditions and types of workloads.     

Optimising a complex discrete event simulation model using a genetic algorithm

A steelworks model is selected as representative of the stochastic and unpredictable behaviour of a complex discrete event simulation model. The steel-works has a number of different entity or object types. Using the number of each entity type as parameters, it is possible to find better and worse combinations of parameters for various management objectives. A simple real-coded genetic algorithm is presented that optimises the parameters, demonstrating the versatility that genetic algorithms offer in solving hard inverse problems.     

Evaluation of carsharing network’s growth strategies through discrete event simulation

Carsharing organizations are nowadays faced with the emergence of new markets due to the growing popularity of their services. To keep up with the growing demand, they have to constantly adapt their network and balance their stations’ capacities by implementing new strategies. These strategies involve creation of new carsharing stations, increasing the capacity of stations, merging or demerging carsharing stations etc. Currently, the decision makers rely on an intuitive strategy selection process which often results in inadequate decisions being made representing an immediate loss in resources, time and market penetration. This paper presents a discrete event simulation based decision support tool that assists the decision makers in selecting best network growth strategies to implement for meeting adequately the demand growth while maximizing the members’ satisfaction level and minimizing the number of vehicles used. Our discrete event simulation model allows modeling the activities at any given set of carsharing stations, regardless of their number and capacities. A benchmarking comparison of different potential strategies is done. An application of the proposed model on a region of Communauto’s Montréal (Québec, Canada) carsharing network is provided.