SIMULATION: Transactions of The Society for Modeling and Simulation International

Mustafee, Navonil; Taylor, Simon J E

doi: 10.1177/0037549713509777pmid: N/A

Zou, Peng; Lü, Ya-shuai; Wu, Ling-da; Chen, Li-li; Yao, Yi-ping

doi: 10.1177/0037549713482026pmid: N/A

Computer simulation is an important tool for studying epidemic dynamics. Owing to the scales and runtime speed requirements, the simulations of large-scale pandemic diseases such as severe acute respiratory syndrome (SARS) and H1N1 influenza usually require high-performance parallel simulation or computation. Previous works on the parallel simulations of large-scale epidemic were implemented on traditional general purpose CPU-based platforms or clusters. As more and more high-performance computation clusters are being built with so-called general-purpose graphics cards, this paper presents the implementation and optimization techniques for social contact network-based large-scale epidemic simulation on GPU clusters. Compared with previous works, this paper focuses on (1) how to efficiently implement the contact network-based parallel epidemic simulation on GPUs, and (2) how to hide communication latencies between processing nodes to improve scalability. Our proposed techniques are implemented and tested on a commodity cluster whose processing nodes are equipped with GPGPUs. The experimental results show that, for the simulation of 20 million individuals and 1.2 billion host contacts on 80 nodes, the execution on GPUs can achieve 7.4×−11.7× speedup over the execution on CPUs.

Hou, Bonan; Yao, Yiping; Wang, Bing; Liao, Dongsheng

doi: 10.1177/0037549713495752pmid: N/A

The modeling and simulation of social networks is an important approach to better understanding complex social phenomena, especially when the inner structure has remarkable impact on behavior. With the availability of unprecedented data sets, simulating large-scale social networks of millions, or even billions, of entities has become a new challenge. Current simulation environments for social studies are mostly sequential and may not be efficient when social networks grow to a certain size. In order to facilitate large-scale social network modeling and simulation, this paper proposes a framework named SUPE-Net, which is based on a parallel discrete event simulation environment YH-SUPE for massively parallel architectures. The framework is designed as a layered architecture with utilities for network generation, algorithms and agent-based modeling. Distributed adjacency lists are used for graph modeling and a reaction–diffusion paradigm is adapted to model dynamical processes. Experiments are performed using PageRank and the susceptible–infected–recovered (SIR) model on social networks with millions of entities. The results demonstrate that SUPE-Net has achieved a speedup of 12, and increased the event-processing rate by 11%, with good scalability and effectiveness.

Zia, Kashif; Farrahi, Katayoun; Riener, Andreas; Ferscha, Alois

doi: 10.1177/0037549713485468pmid: N/A

The simulation of urban mobility is a modeling challenge due to the complexity and scale. The complexity in modeling a social agent is due to three reasons: (i) the agent is behaviorally complex itself due to several interrelated/overlapping modeling aspects; (ii) the setting in which a social agent operates usually demands a multi-resolution approach; and (iii) the consideration of real spatial and population data is the underpinning that has to be realized. In this paper, we propose an agent-based parallel geo-simulation framework of urban mobility based on necessary modeling aspects. The aspect-oriented modeling paradigm relates the models vertically as well as horizontally and highlights the situations requiring multi-resolution interfacing. The framework takes into consideration the importance of technological footprints embedded with social behavior along with essential space and mobility features keeping focus on the importance of the city-scale scenario. We have used a real, high-quality raster map of a medium-sized city in central Europe converting it into a cellular automata (CA). The fine-grained CA readily supports pedestrian mobility and can easily be extended to support other mobility modes. The urban mobility simulation is performed on a real parallel and distributed hardware platform using a CA compatible software platform. Considering city-wide mobility in an emergency scenario, an analysis of the simulation efficiency and agent behavioral response is presented.

Collier, Nicholson; North, Michael

doi: 10.1177/0037549712462620pmid: N/A

In the last decade, agent-based modeling and simulation (ABMS) has been applied to a variety of domains, demonstrating the potential of this technique to advance science, engineering, and policy analysis. However, realizing the full potential of ABMS to find breakthrough research results requires far greater computing capability than is available through current ABMS tools. The Repast for High Performance Computing (Repast HPC) project addresses this need by developing a useful and useable next-generation ABMS system explicitly focusing on larger-scale distributed computing platforms. Repast HPC is intended to smooth the path from small-scale simulations to large-scale distributed simulations through the use of a Logo-like system. This article’s contribution is its detailed presentation of the implementation of Repast HPC as a useful and usable framework, a complete ABMS platform developed explicitly for larger-scale distributed computing systems that leverages modern C++ techniques and the ReLogo language.

Cordasco, Gennaro; De Chiara, Rosario; Mancuso, Ada; Mazzeo, Dario; Scarano, Vittorio; Spagnuolo, Carmine

doi: 10.1177/0037549713489594pmid: N/A

Agent-based simulation models are an increasingly popular tool for research and management in many fields. In executing such simulations “speed” is one of the most general and important issues because of the size and complexity of simulations. But another important issue is the effectiveness of the solution, which consists of how easily usable and portable the solutions are for the users, i.e. the programmers of the distributed simulation. Our study, then, is aimed at efficient and effective distribute simulations by adopting a framework-level approach, with our design and implementation of a framework, D-Mason, which is a parallel version of the Mason library for writing and running simulations of agent-based simulation models. In particular, besides the efficiency due to workload distribution with small overhead, D-Mason at a framework level proves itself effective since it enables the scientists that use the framework (domain expert but with limited knowledge of distributed programming) only minimally aware of the fact that the simulation is running on a distributed environment. Then, we present tests that compare D-Mason against Mason in order to assess the improved scalability and D-Mason capability to exploit heterogeneous distributed hardware. Our tests also show that several massive simulations that are impossible to execute on Mason (e.g. because of CPU and/or memory requirements) can be easily performed using D-Mason.

Pfeifer, Dylan; Valvano, Jonathan; Gerstlauer, Andreas

doi: 10.1177/0037549712472755pmid: N/A

Real-time embedded and cyber-physical systems challenge simulation disciplines due to the heterogeneous tools used to model components in the system exploration and design phases. Termed “heterogeneity,” the mixed-model problem challenges multi-simulator coordination, where event causality must be preserved among simulators with different models of computation, signals, criteria for time advancement, and levels of abstraction. SimConnect and SimTalk enable heterogeneous, distributed, hardware/software co-simulation with a simplified backplane approach, emphasizing the simulation of software interacting with simulated world-model electrical, mechanical, and physical effects. The structure of SimConnect and SimTalk is described, adhering to the properties of a Kahn Process Network. Application of the tools to the coordination of three different simulators (TExaS, Ngspice, and Simulink) is presented to simulate closed-loop, hardware/software-based, Proportional-Integral-Derivative/Pulse-Width-Modulated control of a direct current motor. Results demonstrate agreement among simulator coordinations with configurable tradeoffs in speed versus accuracy.

Li, Zengxiang; Tang, Xueyan; Cai, Wentong; Li, Xiaorong

doi: 10.1177/0037549712470857pmid: N/A

In a distributed virtual environment (DVE), participants located in different places may observe inconsistent views of the simulated virtual world due to message delay and loss in the network. This paper investigates how to compensate for the impact of message delay and loss on consistency in the DVE. We focus on dead reckoning (DR)-based update mechanisms and measure inconsistency by the time–space inconsistency (TSI) metric. We theoretically analyze the TSI of an entity and derive the condition under which the impact of message delay and loss on consistency can be fully compensated for by reducing the DR threshold. Based on the analysis, a compensatory update scheduling algorithm is proposed. Experiments using real traces of a racing car game are conducted to evaluate the compensatory algorithm. The results confirm that if the condition derived in our theoretical analysis is fulfilled, the compensatory algorithm can decrease the TSI of a racing car to the level of the case without message delay and loss when there is sufficient network bandwidth available. Under severe bandwidth constraints, the compensatory algorithm still leads to comparable TSIs of the racing car among the participants regardless of their network conditions so as to enable fair competition.