Farshad Moradi

An agent-based DDM for high level architecture

The Data Distribution Management (DDM) service is one of the six services provided in the Runtime Infrastructure (RTI) of High Level Architecture (HLA). Its purpose is to perform data filtering and reduce irrelevant data communicated between federates. The two DDM schemes proposed for RTI, region based and grid based DDM, are oriented to send as little irrelevant data to subscribers as possible, but only manage to filter part of this information and some irrelevant data is still being communicated. Previously (G. Tan et al., 2000), we employed intelligent agents to perform data filtering in HLA, implemented an agent based DDM in RTI (ARTI) and compared it with the other two filtering mechanisms. The paper reports on additional experiments, results and analysis using two scenarios: the AWACS sensing aircraft simulation and the air traffic control simulation scenario. Experimental results show that compared with other mechanisms, the agent based approach communicates only relevant data and minimizes network communication, and is also comparable in terms of time efficiency. Some guidelines on when the agent based scheme can be used are also given.

A framework for fault-tolerance in HLA-based distributed simulations

The widespread use of simulation in future military systems depends, among others, on the degree of reuse and availability of simulation models. Simulation support in such systems must also cope with failure in software or hardware. Research in fault-tolerant distributed simulation, especially in the context of the high level architecture (HLA), has been quite sparse. Nor does the HLA standard itself cover fault-tolerance extensively. This paper describes a framework, named distributed resource management system (DRMS), for robust execution of federations. The implementation of the framework is based on Web services and semantic Web technology, and provides fundamental services and a consistent mechanism for description of resources managed by the environment. To evaluate the proposed framework, a federation has been developed that utilizes time-warp mechanism for synchronization. In this paper, we describe our approach to fault tolerance and give an example to illustrate how DRMS behaves when it faces faulty federates

Aggregration/disaggregation in hla multi-resolution distributed simulation

Abstract: The High Level Architecture (HLA) is a standard developed by the US Department of Defense (DoD), and is meant to establish interoperability among different types of simulations at multiple locations for the simulation of sometimes highly interactive activities. If there are several HLA compliant federates (simulations) that provide simulation at different levels of resolution, then it would be ideal to combine them into one federation so as to provide realistic training to soldiers of different ranks present within the same federation exercise. The issue of multi-resolution simulation systems thus comes into the picture. Aggregation/disaggregation is used to ensure proper interaction between federates of different levels of resolution. This paper explores the problems of the existing module-based approach to the implementation of aggregation/disaggregation in the HLA environment. Two new approaches and their pros and cons are discussed, and one of them is implemented and tested on a simple dual-resolution federation. A sample simulation run of one of the new approaches is also presented.

Peer-to-Peer-Based Resource Management in Support of HLA-Based Distributed Simulations

In recent years, the concept of peer-to-peer computing has gained renewed interest for sharing resources within and between organizations or individuals. This article describes a decentralized resource management system (DRMS) that uses a network of workstations for the execution and storage of high-level architecture (HLA) federations/federates in a peer-to-peer environment. The implementation of DRMS is based on the open-source project JXTA, which represents an attempt to standardize the peer-to-peer domain. DRMS is part of a Web-based simulation environment supporting collaborative design, development, and execution of HLA federations. This study evaluates the possibilities of using peer-to-peer technology for increasing the reuse and availability of simulation components within the defense modelling and simulation community. More specifically, it addresses the necessary adjustments of simulation components to conform to the requirements of the DRMS and shows that JXTA could provide the foundation for a distributed system that increases the possibilities for reusing simulation components.

Estimating performance of a business process model

We suggest a model for estimating performance of human organizations and business processes. This model is based on subjective assessment of the capabilities of the available human resources, the importance of these capabilities, and the influence of the peripheral factors on the resources. The model can be used to compare different resource allocation schemes in order to choose the most beneficial one. We suggest an extension to Business Process Modeling Notation (BPMN) by including performance measure of performers and the probability by which an outgoing Sequence Flow from a Gateway is chosen. We also propose an analytical method for estimating the overall performance of BPMN in simple cases and a simulation method, which can be used for more complicated scenarios. To illustrate how these methods work, we apply them to part of a military Operational Planning Process and discuss the results.

Component-Based Simulation Model Development Using BOMs and Web Services

Modelling and simulation (M&S) is a valuable tool and provides means for reducing, amongst others, manufacturing and training costs. However, development of simulation models is a multi-disciplinary and time/resource consuming process. Building simulation models through composition and reuse of predefined and already existing validated simulation components is an approach to reduce the associated costs and improving the usability of the models. The base object model, BOM, is a new standard for defining reusable and composable simulation components. However, BOMs lack the necessary expressive power to ensure semantic matching of simulation components. In Web service composition (WSC), composite services are built by assembling existing services in order to address functionalities required by users. In WSC much emphasis has been on including the semantic aspects of the composition, through among others utilization of the semantic Web concept. In this paper we describe a process that has been developed at the Swedish Defence Research Agency (FOI) with the aim to speed up and improve the development of simulation models. This process utilizes the BOM concept and extends it by taking advantage of the techniques used in WSC. We will present our approach and findings based on our implementation of the proposed process

Simulation-based decision support for effects-based planning

In this paper we describe decision support and simulation techniques to facilitate effects-based planning. By using a decision support tool, a decision maker is able to test a number of feasible plans against possible courses of events and decide which of those plans is capable of achieving the desired military end state. The purpose is to evaluate plans and understand their consequences through simulating the events and producing outcomes which result from making alternative decisions. Plans are described in the effects-based approach to operations concept as a set of effects and activities that together will lead to a desired military end state. For each activity we may have several different alternatives. Together they make up all alternative plans, as an activity tree that may be simulated. Simulated plans that are similar in both their structure and consequence are clustered together by a Potts spin neural clustering method. These plans make up a robust set of similar plans that function as ready alternatives should dynamic replanning be necessary as the situation evolves.

Statemachine Matching in BOM Based Model Composition

Base Object Model (BOM) is a component-based standard designed to support reusability and Composability. Reusability helps in reducing time and cost of the development of a simulation process. Composing predefined components such as BOMs is a well known approach to achieve reusability. However, there is a need for a matching mechanism to identify whether a set of components are composable or not. Although BOM provides good model representation, it lacks capability to express semantic and behavioral matching.In this paper we propose an approach for matching behavior of BOM components by matching their statemachines. Our proposed process includes a static and a dynamic matching phase. In the static matching phase, we apply a set of rules to validate the structure of statemachines. In the dynamic matching phase, we execute the statemachines together at an abstract level on our proposed execution framework. We have developed this framework using the State Chart Extensible Markup Language (SCXML), which is a W3C compliant standard. If the execution terminates successfully (i.e. reaches specified final states) we conclude that there is a positive match and the behavior of these BOMs is composable. We describe the matching process and the implementation of our runtime environment in detail and present a case study as proof of concept.

A framework for simulation-based optimization of business process models

The Assignment Problem is a classical problem in the field of combinatorial optimization, having a wide range of applications in a variety of contexts. In general terms, the Assignment Problem consists of determining the best assignment of tasks to agents according to a predefined objective function. Different variants of the Assignment Problem have been extensively investigated in the literature in the last 50 years. In this work, we introduce and analyze the problem of optimizing a business process model with the objective of finding the most beneficial assignment of tasks to agents. Despite similarities, this problem is distinguished from the traditional Assignment Problem in that we consider tasks to be part of a business process model, being interconnected according to defined rules and constraints. In other words, assigning a business process to agents is a more complex form of the Assignment Problem. Two main categories of business processes, assignment-independent and assignment-dependent, are distinguished. In the first category, different assignments of tasks to agents do not affect the flow of the business process, while processes in the second category contain critical tasks that may change the workflow, depending on who performs them. In each category several types of processes are studied. Algorithms for finding optimal and near-optimal solutions to these categories are presented. For the first category, depending on the type of process, the Hungarian algorithm is combined with either the analytical method or simulation to provide an optimal solution. For the second category, we introduce two algorithms. The first one finds an optimal solution, but is feasible only when the number of critical tasks is small. The second algorithm is applicable to large number of critical tasks, but provides a near-optimal solution. In the second algorithm a hill-climbing heuristic method is combined with the Hungarian algorithm and simulation to find an overall near-optimal solution. A series of tests is conducted which demonstrates that the proposed algorithms efficiently find optimal solutions for assignment-independent and near-optimal solutions for assignment-dependent processes.

SDR: A Semantic Based Distributed Repository for Simulation Models and Resources

Recent advances in Internet, peer-to-peer and grid technologies have made collaboration and resource sharing across organizational boundaries more feasible. Today, it is essential for many organizations to be able to discover, share and manage distributed resources in a transparent, meaningful and secure way. A fundamental problem is locating, matching and composing resources or services of interest. In this paper we describe our initial work designing and developing a semantic based distributed repository for secure sharing of simulation models, components and related resources such as computer resources. We propose an overlay architecture which combines advances in semantic Web, peer-to-peer and grid techniques. In our project at the Swedish Defence Research Agency (FOI) we had a need for a repository of simulation related resources and having identified our requirements we found that there was no suitable of-the-shelf system available. We describe the design, tools and a prototype implementation of this system - the semantic based distributed repository (SDR) and conclude with our experiences and some raised issues. We argue that although some of the used techniques still are a bit immature and need further improvements a system like the SDR has a lot of potential and can also be used in other domains than modeling and simulation