John Daniel Siirola

Toward agent-based process systems engineering: proposed framework and application to non-convex optimization

Agent-based computer systems are surprisingly effective at solving complex problems. Built by combining autonomous computer routines, or agents, with low-bandwidth communication capabilities, these systems typically perform significantly better than the individual routines operating alone. One source of this improvement lies in the cooperative collaboration among the individual agents that compose the system. This work proposes a modular framework for implementing agent-based systems for engineering design. Using a variety of different algorithmic agents, the key ideas are highlighted by identifying multiple identical global optima for a non-convex optimization problem with numerous local minima. The results show that inter- and intra-agent collaboration have a significant impact on system performance. Further, the system can be parallelized with little or no penalty. By observing and analyzing the interactions among individual agents, we gain insights that will aid in the development and management of a conceptual design system for truly hard and large problems.

Evaluating Moving Target Defense with PLADD

This project evaluates the effectiveness of moving target defense (MTD) techniques using a new game we have designed, called PLADD, inspired by the game FlipIt [28]. PLADD extends FlipIt by incorporating what we believe are key MTD concepts. We have analyzed PLADD and proven the existence of a defender strategy that pushes a rational attacker out of the game, demonstrated how limited the strategies available to an attacker are in PLADD, and derived analytic expressions for the expected utility of the game’s players in multiple game variants. We have created an algorithm for finding a defender’s optimal PLADD strategy. We show that in the special case of achieving deterrence in PLADD, MTD is not always cost effective and that its optimal deployment may shift abruptly from not using MTD at all to using it as aggressively as possible. We believe our effort provides basic, fundamental insights into the use of MTD, but conclude that a truly practical analysis requires model selection and calibration based on real scenarios and empirical data. We propose several avenues for further inquiry, including (1) agents with adaptive capabilities more reflective of real world adversaries, (2) the presence of multiple, heterogeneous adversaries, (3) computational game theory-based approaches such as coevolution to allow scaling to the real world beyond the limitations of analytical analysis and classical game theory, (4) mapping the game to real-world scenarios, (5) taking player risk into account when designing a strategy (in addition to expected payoff), (6) improving our understanding of the dynamic nature of MTD-inspired games by using a martingale representation, defensive forecasting, and techniques from signal processing, and (7) using adversarial games to develop inherently resilient cyber systems.

Modeling Mathematical Programs with Equilibrium Constraints in Pyomo

We describe new capabilities for modeling MPEC problems within the Pyomo modeling software. These capabilities include new modeling components that represent complementar- ity conditions, modeling transformations for re-expressing models with complementarity con- ditions in other forms, and meta-solvers that apply transformations and numeric optimization solvers to optimize MPEC problems. We illustrate the breadth of Pyomos modeling capabil- ities for MPEC problems, and we describe how Pyomos meta-solvers can perform local and global optimization of MPEC problems.

Modeling and Optimization of Superstructure-based Stochastic Programs for Risk-aware Decision Support

Abstract This manuscript presents a unified software framework for modeling and optimizing large-scale engineered systems with uncertainty. We propose a Python-based “block-oriented” modeling approach for representing the discrete components within the system. Through the use of a modeling components library, the block-oriented approach facilitates a clean separation of system superstructure from the details of individual components. This approach also lends itself naturally to expressing design and operational decisions as disjunctive expressions over the component blocks. We then apply a Python-based risk and uncertainty analysis library that leverages the explicit representation of the mathematical program in Python to automatically expand the deterministic system model into a multi-stage stochastic program, which can then be solved either directly or via decomposition-based solution strategies. This manuscript demonstrates the application of this modeling approach for risk-aware analysis of an electric distribution system.

RPI : A remote process interface library for distributed clusters

Abstract A fundamental issue in working with and designing software applications for distributed computer clusters is selecting the mechanism for providing inter-process and inter-computer communication and coordination. Although several well-established facilities exist for managing access to remote computers, there are applications where these facilities do not adequately meet the application needs. In this paper, we summarize the characteristics of these applications and present a new software package – the remote process interface (RPI) library – as a viable alternative for managing independent processes executing on remote computers.

A framework for modeling and optimizing dynamic systems under uncertainty

Abstract Algebraic modeling languages (AMLs) have drastically simplified the implementation of algebraic optimization problems. However, there are still many classes of optimization problems that are not easily represented in most AMLs. These classes of problems are typically reformulated before implementation, which requires significant effort and time from the modeler and obscures the original problem structure or context. In this work we demonstrate how the Pyomo AML can be used to represent complex optimization problems using high-level modeling constructs. We focus on the operation of dynamic systems under uncertainty and demonstrate the combination of Pyomo extensions for dynamic optimization and stochastic programming. We use a dynamic semibatch reactor model and a large-scale bubbling fluidized bed adsorber model as test cases.

Optimization of large-scale heterogeneous system-of-systems models.

Decision makers increasingly rely on large-scale computational models to simulate and analyze complex man-made systems. For example, computational models of national infrastructures are being used to inform government policy, assess economic and national security risks, evaluate infrastructure interdependencies, and plan for the growth and evolution of infrastructure capabilities. A major challenge for decision makers is the analysis of national-scale models that are composed of interacting systems: effective integration of system models is difficult, there are many parameters to analyze in these systems, and fundamental modeling uncertainties complicate analysis. This project is developing optimization methods to effectively represent and analyze large-scale heterogeneous system of systems (HSoS) models, which have emerged as a promising approach for describing such complex man-made systems. These optimization methods enable decision makers to predict future system behavior, manage system risk, assess tradeoffs between system criteria, and identify critical modeling uncertainties.