Search

Simulation models are becoming an increasingly popular tool for the analysis and optimization of complex real systems in different fields. Finding an optimal system design requires performing a large sweep over the parameter space in an organized way. Hence, the model optimization process is extremely demanding from a computational point of view, as it requires careful, time-consuming, complex orchestration of coordinated executions. In this paper, we present the design of SOF (Simulation Optimization and exploration Framework in the cloud), a framework which exploits the computing power of a cloud computational environment in order to carry out effective and efficient simulation optimization strategies. SOF offers several attractive features. Firstly, SOF requires "zero configuration", as it does not require any additional software installed on the remote node; only standard Apache Hadoop and SSH access are sufficient. Secondly, SOF is transparent to the user, since the user is totally unaware that the system operates on a distributed environment. Finally, SOF is highly customizable and programmable, since it enables the running of different simulation optimization scenarios using diverse programming languages - provided that the hosting platform supports them - and different simulation toolkits, as developed by the modeler. The tool has been fully developed and is available on a public repository.11SOF GitHub public repository, https://github.com/isislab-unisa/sof. under the terms of the open source Apache License. It has been tested and validated on several private platforms, such as a dedicated cluster of workstations, as well as on public platforms, including the Hortonworks Data Platform and Amazon Web Services Elastic MapReduce solution.

Scientific Computing (SC) is a multidisciplinary field that uses the computational approach to understand and study complex artificial and natural systems belonging many scientific sectors. Optimization via Simulation (OvS) is a fast developing area in SC field. OvS combines classical optimization algorithms and stochastic simulations to face problems with unknown and/or dynamic data distribution. We present Heterogeneous Simulation Optimization (HSO), an architecture that enable to distribute the OvS process on an Heterogeneous Computing systems. HSO is designed according to two levels of heterogeneity: hardware heterogeneity, that is the ability to exploit the computational power of several general-purpose CPUs and/or hardware accelerators such as Graphics Processing Units (GPUs); programming languages heterogeneity, that is the capability to develop the OvS methodology combining different programming languages such as C++, C, Clojure, Erlang, Go, Haskel, Java, Node.js, Objective-C, PHP, Python, Scala and many others. The proposed HSO architecture has been fully developed and is available on a public GitHub repository. We have validated and tested the scalability of HSO developing two different use cases that show both the levels of heterogeneity, and showing how to exploit Optimal Computing Budget Allocation (OCBA) algorithm and a Genetic Algorithm in a OvS process.

Computational Social Science (CSS) involves interdisciplinary fields and exploits computational methods, such as social network analysis as well as computer simulation with the goal of better understanding social phenomena. Agent-Based Models (ABMs) represent an effective research tool for CSS and consist of a class of models, which, aim to emulate or predict complex phenomena through a set of simple rules (i.e., independent actions, interactions and adaptation), performed by multiple agents. The efficiency and scalability of ABMs systems are typically obtained distributing the overall computation on several machines, which interact with each other in order to simulate a specific model. Unfortunately, the design of a distributed simulation model is particularly challenging, especially for domain experts who sporadically are computer scientists and are not used to developing parallel code. D-MASON framework is a distributed version of the MASON library for designing and executing ABMs in a distributed environment ensuring scalability and easiness. D-MASON enable the developer to exploit the computing power of distributed environment in a transparent manner; the developer has to do simple incremental modifications to existing MASON models, without re-designing them. This paper presents several novel features and architectural improvements introduced in the D-MASON framework: an improved space partitioning strategy, a distributed 3D field, a distributed network field, a decentralized communication layer, a novel memory consistency mechanism and the integration to cloud environments. Full documentation, additional tutorials, and other material can be found at https://github.com/isislab-unisa/dmason where the framework can be downloaded.

Agent-based simulations are becoming widespread among scientists from different areas, who use them to model increasingly complex problems. To cope with the growing computational complexity, parallel and distributed implementations have been developed for a wide range of platforms. However, it is difficult to have simulations that are portable to different platforms while still achieving high performance. We present OpenABL, a domain-specific language for portable, high-performance, parallel agent modeling. It comprises an easy-to-program language that relies on high-level abstractions for programmability and explicitly exploits agent parallelism to deliver high performance. A source-to-source compiler translates the input code to a high-level intermediate representation exposing parallelism, locality and synchronization, and, thanks to an architecture based on pluggable backends, generates target code for multi-core CPUs, GPUs, large clusters and cloud systems. OpenABL has been evaluated on six applications from various fields such as ecology, animation, and social sciences. The generated code scales to large clusters and performs similarly to hand-written target-specific code, while requiring significantly fewer lines of codes.

Work partitioning is a key challenge with ap- plications in many scientific and technological fields. The problem is very well studied with a rich literature on both distributed and parallel computing architectures. In this paper we deal with the work partitioning problem for parallel and distributed agent-based simulations which aims at (i) balancing the overall load distribution, (ii) minimizing, at the same time, the communication overhead due to agents' inter-dependencies. We introduce a classification taxonomy of work partitioning strategies and present a space-based work partitioning ap- proach, based on a Quad-tree data structure, which enables to: identify a good space partitioning (even when the distribution of agents on the fields is non-uniform) with a limited impact in terms of communication. Being a multi-objective problem, the results are difficult to compare and it is hard to foresee what can be the impact of one solution. For this reason we evaluate our strategy in a real context using a well-known behavior (the boids flocking model), on a distributed agent based simulation framework (D-MASON). The results show that our proposal provides a sensible impact on the performances of the system and scales in terms of the number of logical processors.

Simulation models are becoming an increasingly popular tool for the analysis and optimization of complex real systems in different fields. Finding an optimal system design requires performing a large parameter sweep. In this paper, we present the design of SOF (Simulation Optimization and exploration Framework on the cloud), a framework which exploits the computing power of a cloud computational environment in order to realize effective and efficient simulation optimization strategies. SOF offers several attractive features: SOF requires "zero configuration" as it does not require any additional software installed on the remote node, SOF is transparent to the user, since the user is totally unaware that system operates on a distributed environment, SOF is highly customizable and programmable, since it enables the running of different simulation optimization scenarios on different simulation toolkits. The tool has been fully developed and is available on a public repository under the Apache public licence.

Agent-Based Models (ABMs) denote a class of models which, by simulating the behavior of multiple agents (i.e., independent actions, interactions and adaptation), aim to emulate and/or predict complex phenomena. One of the general features of ABM simulations is their experimental capacity, that requires a viable and reliable infrastructure to interact with a running simulation, monitoring its behaviour, as it proceeds, and applying changes to the configurations at run time, in order to study "what if" scenarios. A common approach for improving the efficiency and the effectiveness of ABMs as a research tool is to distribute the overall computation on a number of machines, which makes the design of the simulation model particularly challenging. D-MASON framework is a distributed version of the MASON library for writing and running Agent-based simulations. We briefly present D-MASON architecture and functionalities. Then we presents its novel features: a distributed network field and a novel communication layer dedicated to massive parallel machines. The main contribution of the paper is in providing a memory consistency modeling, where the previous state of theagent is made available (consistently) for all other agents (even the one on other processors) and this is obtained by exploiting the Java Method Handler mechanism. Full documentation, additional tutorials and other material can be found at www.dmason.org where the framework can be downloaded.