ACM Transactions on Modeling and Computer Simulation最新文献

Fitting stochastic input-process models to data and then sampling from them are key steps in a simulation study, but highly challenging to non-experts. We present Neural Input Modeling (NIM), a generative-neural-network (GNN) framework that exploits modern data-rich environments to automatically capture simulation input processes and then generate samples from them. The basic GNN that we develop, called NIM-VL, comprises (i) a variational-autoencoder (VAE) architecture that learns the probability distribution of the input data while avoiding overfitting and (ii) Long Short-Term Memory (LSTM) components that concisely capture statistical dependencies across time. We show how the basic GNN architecture can be modified to exploit known distributional properties—such as i.i.d. structure, nonnegativity, and multimodality—in order to increase accuracy and speed, as well as to handle multivariate processes, categorical-valued processes, and extrapolation beyond the training data for certain nonstationary processes. We also introduce an extension to NIM called “conditional” NIM (CNIM), which can learn from training data obtained under various realizations of a (possibly time-series-valued) stochastic “condition”, such as temperature or inflation rate, and then generate sample paths given a value of the condition not seen in the training data. This enables users to simulate a system under a specific working condition by customizing a pre-trained model; CNIM also facilitates what-if analysis. Extensive experiments show the efficacy of our approach. NIM can thus help overcome one of the key barriers to simulation for non-experts.

We evaluate a stochastic upper bound on the response time Probability Density Function (PDF) of complex workflows through an efficient and accurate compositional approach. Workflows consist of activities having generally distributed stochastic durations with bounded supports, composed through sequence, choice/merge, and balanced/unbalanced split/join operators, possibly breaking the structure of well-formed nesting. Workflows are specified using a formalism defined in terms of Stochastic Time Petri Nets (STPNs), that permits decomposition into a hierarchy of subworkflows with positively correlated response times, guaranteeing that a stochastically larger end-to-end response time PDF is obtained when intermediate results are approximated by stochastically larger PDFs and when dependencies are simplified by replicating activities appearing in multiple subworkflows. In particular, an accurate stochastically larger PDF is obtained by combining shifted truncated Exponential terms with positive or negative rates. Experiments are performed on sets of manually and randomly generated models with increasing complexity, illustrating under which conditions different decomposition heuristics work well in terms of accuracy and complexity, and showing that the proposed approach outperforms simulation having the same execution time.

We propose a new framework of a neural network-assisted sequential structured simulator to model, estimate, and simulate a wide class of sequentially generated data. Neural networks are integrated into the sequentially structured simulators in order to capture potential nonlinear and complicated sequential structures. Given representative real data, the neural network parameters in the simulator are estimated and calibrated through a Wasserstein training process, without restrictive distributional assumptions. The target of Wasserstein training is to enforce the joint distribution of the simulated data to match the joint distribution of the real data in terms of Wasserstein distance. Moreover, the neural network-assisted sequential structured simulator can flexibly incorporate various kinds of elementary randomness and generate distributions with certain properties such as heavy-tail, without the need to redesign the estimation and training procedures. Further, regarding statistical properties, we provide results on consistency and convergence rate for the estimation procedure of the proposed simulator, which are the first set of results that allow the training data samples to be correlated. We then present numerical experiments with synthetic and real data sets to illustrate the performance of the proposed simulator and estimation procedure.

We derive new discrete event simulation algorithms for marked time point processes. The main idea is to couple a special structure, namely the associated local independence graph, as defined by Didelez, with the activity tracking algorithm of Muzy for achieving high-performance asynchronous simulations. With respect to classical algorithms, this allows us to drastically reduce the computational complexity, especially when the graph is sparse.

Adaptive Monte Carlo variance reduction is an effective framework for running a Monte Carlo simulation along with a parameter search algorithm for variance reduction, whereas an initialization step is required for preparing problem parameters in some instances. In spite of the effectiveness of adaptive variance reduction in various fields of application, the length of the preliminary phase has often been left unspecified for the user to determine on a case-by-case basis, much like in typical sequential frameworks. This uncertain element may possibly be even fatal in realistic finite-budget situations, since the pilot run may take most of the budget, or possibly use up all of it. To unnecessitate such an ad hoc initialization step, we develop a batching procedure in adaptive variance reduction, and provide an implementable formula of the learning rate in the parameter search which minimizes an upper bound of the theoretical variance of the empirical batch mean. We analyze decay rates of the minimized upper bound towards the minimal estimator variance with respect to the predetermined computing budget, and provide convergence results as the computing budget increases progressively when the batch size is fixed. Numerical examples are provided to support theoretical findings and illustrate the effectiveness of the proposed batching procedure.

The rapid introduction of mobile navigation aides that use real-time road network information to suggest alternate routes to drivers is making it more difficult for researchers and government transportation agencies to understand and predict the dynamics of congested transportation systems. Computer simulation is a key capability for these organizations to analyze hypothetical scenarios; however, the complexity of transportation systems makes it challenging for them to simulate very large geographical regions, such as multi-city metropolitan areas. In this article, we describe enhancements to the Mobiliti parallel traffic simulator to model dynamic rerouting behavior with the addition of vehicle controller actors and vehicle-to-controller reroute requests. The simulator is designed to support distributed-memory parallel execution using discrete event simulation and be scalable on high-performance computing platforms. We demonstrate the potential of the simulator by analyzing the impact of varying the population penetration rate of dynamic rerouting on the San Francisco Bay Area road network. Using high-performance parallel computing, we can simulate a day in the San Francisco Bay Area with 19 million vehicle trips with 50 percent dynamic rerouting penetration over a road network with 0.5 million nodes and 1 million links in less than three minutes. We present a sensitivity study on the dynamic rerouting parameters, discuss the simulator’s parallel scalability, and analyze system-level impacts of changing the dynamic rerouting penetration. Furthermore, we examine the varying effects on different functional classes and geographical regions and present a validation of the simulation results compared to real-world data.

Building performance models for software services in DevOps is costly and error-prone. Accurate service demand distribution estimation is critical to precisely modeling queueing behaviors and performance prediction. However, current estimation methods focus on capturing the mean service demand, disregarding higher-order moments of the distribution that still can largely affect prediction accuracy. To address this limitation, we propose to estimate higher moments of the service demand distribution for a microservice from monitoring traces. We first generate a closed queueing model to abstract software performance and use it to model the departure process of requests completed by the software service as a Markovian arrival process (MAP). This allows formulating the estimation of service demand into an optimization problem, which aims to find the first multiple moments of the service demand distribution that maximize the likelihood of the MAP using generated the measured inter-departure times. We then estimate the service demand distribution for different classes of service with a maximum likelihood algorithm and novel heuristics to mitigate the computational cost of the optimization process for scalability. We apply our method to real traces from a microservice-based application and demonstrate that its estimations lead to greater prediction accuracy than exponential distributions assumed in traditional service demand estimation approaches for software services.