Decomposition theory for multidisciplinary design optimization problems with mixed integer quasiseparable subsystems

Numerous hierarchical and nonhierarchical decomposition strategies for the optimization of large scale systems, comprised of interacting subsystems, have been proposed. With a few exceptions, all of these strategies are essentially heuristic in nature. Recent work considered a class of optimization problems, called quasiseparable, narrow enough for a rigorous decomposition theory, yet general enough to encompass many large scale engineering design problems. The subsystems for these problems involve local design variables and global system variables, but no variables from other subsystems. The objective function is a sum of a global system criterion and the subsystems’ criteria. The essential idea is to give each subsystem a budget and global system variable values, and then ask the subsystems to independently maximize their constraint margins. Using these constraint margins, a system optimization then adjusts the values of the system variables and subsystem budgets. The subsystem margin problems are totally independent, always feasible, and could even be done asynchronously in a parallel computing context. An important detail is that the subsystem tasks, in practice, would be to construct response surface approximations to the constraint margin functions, and the system level optimization would use these margin surrogate functions. The present paper extends the quasiseparable necessary conditions for continuous variables to include discrete subsystem variables, although the continuous necessary and sufficient conditions do not extend to include integer variables.     

An artificial immune network for multiobjective optimization problems

Multiobjective optimization is an important problem of great complexity and evolutionary algorithms have been established as a dominant approach in the field. This article suggests a method for approximating the Pareto front of a given function based on artificial immune networks. The proposed method uses cloning and mutation on a population of antibodies to create local subsets of the Pareto front. Elements of these local fronts are combined, in a way that maximizes diversity, to form the complete Pareto front of the function. The method is tested on a number of well-known benchmark problems, as well as an engineering problem. Its performance is compared against state-of-the-art algorithms, yielding promising results.     

Interactive solution to multiobjective optimization problems

Sensitivity analysis is presented as a natural addition to interactive multiobjective optimization methods based on compromise programming. It is shown that useful information regarding trade-offs in the objectives can be generated effectively by means of an analysis of the sensitivity of solutions to variations in preference structures. An implementation based on sequential quadrative programming is provided and examples are given for illustration.     

Ant algorithm for single and multiobjective reliability optimization problems

The paper illustrates the application of the ant colony optimization algorithm to solve both continuous function and combinatorial optimization problems in reliability engineering. The ant algorithm is combined with the strength Pareto fitness assignment procedure to handle multiobjective problems. Further, a clustering procedure has been applied to prune the Pareto set and to maintain diversity. Benchmark case examples show the superiority of the ant algorithm to such problems. Copyright © 2002 John Wiley & Sons, Ltd.     

Single- and multiobjective optimization problems in robust parameter design

This paper reviews the evolution of off-line quality engineering methods with respect to one or more quality criteria, and presents some recent results. The fundamental premises that justify the use of robust product/process design are established with an illustrative example. The use of designed experiments to model quality criteria and their optimization is briefly reviewed. The fact that most design-for-quality problems involve multiple quality criteria motivates the development of multiobjective optimization techniques for robust parameter design. Two situations are considered: one in which response surface models for the quality characteristics can be obtained using regression and considered over a continuous factor space, and one in which the problem scenario and the experiment permit only discrete parameter settings for the design factors. In the former scenario, a multiobjective optimization technique based on the reference-point method is presented; this technique also incorporates an inference mechanism to deal with uncertainty in the response surface models caused by finite, noisy data. In the discrete-factors scenario, an efficient method to reduce computational complexity for a class of models is presented.     

Global formulation for interactive multiobjective optimization

Interactive methods are useful and realistic multiobjective optimization techniques and, thus, many such methods exist. However, they have two important drawbacks when using them in real applications. Firstly, the question of which method should be chosen is not trivial. Secondly, there are rather few practical implementations of the methods. We introduce a general formulation that can accommodate several interactive methods. This provides a comfortable implementation framework for a general interactive system. Besides, this implementation allows the decision maker to choose how to give preference information to the system, and enables changing it anytime during the solution process. This change-of-method option provides a very flexible framework for the decision maker.     

Quadratic scalarization for decomposed multiobjective optimization

Practical applications in multidisciplinary engineering design, business management, and military planning require distributed solution approaches for solving nonconvex, multiobjective optimization problems (MOPs). Under this motivation, a quadratic scalarization method (QSM) is developed with the goal to preserve decomposable structures of the MOP while addressing nonconvexity in a manner that avoids a high degree of nonlinearity and the introduction of additional nonsmoothness. Under certain assumptions, necessary and sufficient conditions for QSM-generated solutions to be weakly and properly efficient for an MOP are developed, with any form of efficiency being understood in a local sense. QSM is shown to correspond with the relaxed, reformulated weighted-Chebyshev method as a special case. An example is provided for demonstrating the application of QSM to a nonconvex MOP.     

Interactive multiobjective optimization with NIMBUS for decision making under uncertainty

We propose an interactive method for decision making under uncertainty, where uncertainty is related to the lack of understanding about consequences of actions. Such situations are typical, for example, in design problems, where a decision maker has to make a decision about a design at a certain moment of time even though the actual consequences of this decision can be possibly seen only many years later. To overcome the difficulty of predicting future events when no probabilities of events are available, our method utilizes groupings of objectives or scenarios to capture different types of future events. Each scenario is modeled as a multiobjective optimization problem to represent different and conflicting objectives associated with the scenarios. We utilize the interactive classification-based multiobjective optimization method NIMBUS for assessing the relative optimality of the current solution in different scenarios. This information can be utilized when considering the next step of the overall solution process. Decision making is performed by giving special attention to individual scenarios. We demonstrate our method with an example in portfolio optimization.     

Bi-level optimization for a dynamic multiobjective problem

In this article, a bi-level optimization problem covering upper (design) and lower (operation) levels is defined and a solution procedure for bi-level optimization problems is presented. This is devised as a dynamic multiobjective optimization problem, i.e. the values of the control and state variables change over a predefined time horizon and several competing criteria are optimized simultaneously. Moreover, the interaction between the upper and lower levels is analysed. The benefits of bi-level dynamic multiobjective optimization are illustrated in detail by examining an industrial case in which the design of a paper mill (upper level) and the mill operation (lower level) are optimized at the same time. However, the problem definition and the solution procedure are not limited to any specific application but can be exploited in many different industrial areas.     

Pareto navigator for interactive nonlinear multiobjective optimization

We describe a new interactive learning-oriented method called Pareto navigator for nonlinear multiobjective optimization. In the method, first a polyhedral approximation of the Pareto optimal set is formed in the objective function space using a relatively small set of Pareto optimal solutions representing the Pareto optimal set. Then the decision maker can navigate around the polyhedral approximation and direct the search for promising regions where the most preferred solution could be located. In this way, the decision maker can learn about the interdependencies between the conflicting objectives and possibly adjust one’s preferences. Once an interesting region has been identified, the polyhedral approximation can be made more accurate in that region or the decision maker can ask for the closest counterpart in the actual Pareto optimal set. If desired, (s)he can continue with another interactive method from the solution obtained. Pareto navigator can be seen as a nonlinear extension of the linear Pareto race method. After the representative set of Pareto optimal solutions has been generated, Pareto navigator is computationally efficient because the computations are performed in the polyhedral approximation and for that reason function evaluations of the actual objective functions are not needed. Thus, the method is well suited especially for problems with computationally costly functions. Furthermore, thanks to the visualization technique used, the method is applicable also for problems with three or more objective functions, and in fact it is best suited for such problems. After introducing the method in more detail, we illustrate it and the underlying ideas with an example.     

Efficient dynamic resampling for dominance-based multiobjective evolutionary optimization

Multi-objective optimization problems are often subject to the presence of objectives that require expensive resampling for their computation. This is the case for many robustness metrics, which are frequently used as an additional objective that accounts for the reliability of specific sections of the solution space. Typical robustness measurements use resampling, but the number of samples that constitute a precise dispersion measure has a potentially large impact on the computational cost of an algorithm. This article proposes the integration of dominance based statistical testing methods as part of the selection mechanism of evolutionary multi-objective genetic algorithms with the aim of reducing the number of fitness evaluations. The performance of the approach is tested on five classical benchmark functions integrating it into two well-known algorithms, NSGA-II and SPEA2. The experimental results show a significant reduction in the number of fitness evaluations while, at the same time, maintaining the quality of the solutions.     

Trade-off analysis approach for interactive nonlinear multiobjective optimization

When solving multiobjective optimization problems, there is typically a decision maker (DM) who is responsible for determining the most preferred Pareto optimal solution based on his preferences. To gain confidence that the decisions to be made are the right ones for the DM, it is important to understand the trade-offs related to different Pareto optimal solutions. We first propose a trade-off analysis approach that can be connected to various multiobjective optimization methods utilizing a certain type of scalarization to produce Pareto optimal solutions. With this approach, the DM can conveniently learn about local trade-offs between the conflicting objectives and judge whether they are acceptable. The approach is based on an idea where the DM is able to make small changes in the components of a selected Pareto optimal objective vector. The resulting vector is treated as a reference point which is then projected to the tangent hyperplane of the Pareto optimal set located at the Pareto optimal solution selected. The obtained approximate Pareto optimal solutions can be used to study trade-off information. The approach is especially useful when trade-off analysis must be carried out without increasing computation workload. We demonstrate the usage of the approach through an academic example problem.     

Collaborative optimization with inverse reliability for multidisciplinary systems uncertainty analysis

This article introduces a method which combines the collaborative optimization framework and the inverse reliability strategy to assess the uncertainty encountered in the multidisciplinary design process. This method conducts the sub-system analysis and optimization concurrently and then improves the process of searching for the most probable point (MPP). It reduces the load of the system-level optimizer significantly. This advantage is specifically more prominent for large-scale engineering system design. Meanwhile, because the disciplinary analyses are treated as the equality constraints in the disciplinary optimization, the computation load can be further reduced. Examples are used to illustrate the accuracy and efficiency of the proposed method.     

Riemannian optimization and multidisciplinary design optimization

Riemannian Optimization (RO) generalizes standard optimization methods from Euclidean spaces to Riemannian manifolds. Multidisciplinary Design Optimization (MDO) problems exist on Riemannian manifolds, and with the differential geometry framework which we have previously developed, we can now apply RO techniques to MDO. Here, we provide background theory and a literature review for RO and give the necessary formulae to implement the Steepest Descent Method (SDM), Newton’s Method (NM), and the Conjugate Gradient Method (CGM), in Riemannian form, on MDO problems. We then compare the performance of the Riemannian and Euclidean SDM, NM, and CGM algorithms on several test problems (including a satellite design problem from the MDO literature); we use a calculated step size, line search, and geodesic search in our comparisons. With the framework’s induced metric, the RO algorithms are generally not as effective as their Euclidean counterparts, and line search is consistently better than geodesic search. In our post-experimental analysis, we also show how the optimization trajectories for the Riemannian SDM and CGM relate to design coupling and thereby provide some explanation for the observed optimization behaviour. This work is only a first step in applying RO to MDO, however, and the use of quasi-Newton methods and different metrics should be explored in future research.     

Usage scenarios for design space exploration with a dynamic multiobjective optimization formulation

In a recent publication, we presented a new strategy for engineering design and optimization, which we termed formulation space exploration. The formulation space for an optimization problem is the union of all variable and design objective spaces identified by the designer as being valid and pragmatic problem formulations. By extending a computational search into this new space, the solution to any optimization problem is no longer predefined by the optimization problem formulation. This method allows a designer to both diverge the design space during conceptual design and converge onto a solution as more information about the design objectives and constraints becomes available. Additionally, we introduced a new way to formulate multiobjective optimization problems, allowing the designer to change and update design objectives, constraints, and variables in a simple, fluid manner that promotes exploration. In this paper, we investigate three usage scenarios where formulation space exploration can be utilized in the early stages of design when it is possible to make the greatest contributions to development projects. Specifically, we look at formulation space boundary exploration, Pareto frontier generation for multiple concepts in the formulation space, and a new way to perform targeted boundary expansion. The benefits of these methods are illustrated with the conceptual design of an impact driver.     

On scalarizing functions in multiobjective optimization

Scalarizing functions play an essential role in solving multiobjective optimization problems. Many different scalarizing functions have been suggested in the literature based on different approaches. Here we concentrate on classification and reference point-based functions. We present a collection of functions that have been used in interactive methods as well as some modifications. We compare their theoretical properties and numerical behaviour. In particular, we are interested in the relation between the information provided and the results obtained. Our aim is to select some of them to be used in our WWW-NIMBUS optimization system. Received: June 11, 2001 / Accepted: October 31, 2001     

Computational strategy for multiobjective optimization of composite stiffened panels

This paper presents a multiobjective optimization methodology for composite stiffened panels. The purpose is to improve the performances of an existing design of stiffened composite panels in terms of both its first buckling load and ultimate collapse or failure loads. The design variables are the stacking sequences of the skin and of the stiffeners of the panel. The optimization is performed using a multiobjective evolutionary algorithm specifically developed for the design of laminated parts. The algorithm takes into account the industrial design guidelines for stacking sequence design. An original method is proposed for the initialization of the optimization that significantly accelerates the search for the Pareto front. In order to reduce the calculation time, Radial Basis Functions under Tension are used to approximate the objective functions. Special attention is paid to generalization errors around the optimum. The multiobjective optimization results in a wide set of trade-offs, offering important improvements for both considered objectives, among which the designer can make a choice.     

A computationally efficient metamodeling approach for expensive multiobjective optimization

This paper explores a new metamodeling framework that may collapse the computational explosion that characterizes the modeling of complex systems under a multiobjective and/or multidisciplinary setting. Under the new framework, a pseudo response surface is constructed for each design objective for each discipline. This pseudo response surface has the unique property of being highly accurate in Pareto optimal regions, while it is intentionally allowed to be inaccurate in other regions. In short, the response surface for each design objective is accurate only where it matters. Because the pseudo response surface is allowed to be inaccurate in other regions of the design space, the computational cost of constructing it is dramatically reduced. An important distinguishing feature of the new framework is that the response surfaces for all the design objectives are constructed simultaneously in a mutually dependent fashion, in a way that identifies Pareto regions for the multiobjective problem. The new framework supports the puzzling notion that it is possible to obtain more accuracy and radically more design space exploration capability, while actually reducing the computation effort. This counterintuitive metamodeling paradigm shift holds the potential for identifying highly competitive products and systems that are well beyond today’s state of the art.     

Minimax multiobjective optimization in structural design

The use of multiobjective optimization techniques, which may be regarded as a systematic sensitivity analysis, for the selection and modification of system parameters is presented. A minimax multiobjective optimization model for structural optimization is proposed. Three typical multiobjective optimization techniques—goal programming, compromise programming and the surrogate worth trade-off method—are used to solve such a problem. The application of multiobjective optimization techniques to the selection of system parameters and large scale structural design optimization problems is the main purpose of this paper.     

Handling constraints using multiobjective optimization concepts

In this paper, we propose a new constraint‐handling technique for evolutionary algorithms which we call inverted‐shrinkable PAES (IS‐PAES). This approach combines the use of multiobjective optimization concepts with a mechanism that focuses the search effort onto specific areas of the feasible region by shrinking the constrained search space. IS‐PAES also uses an adaptive grid to store the solutions found, but has a more efficient memory‐management scheme than its ancestor (the Pareto archived evolution strategy for multiobjective optimization). The proposed approach is validated using several examples taken from the standard evolutionary and engineering optimization literature. Comparisons are provided with respect to the stochastic ranking method (one of the most competitive constraint‐handling approaches used with evolutionary algorithms currently available) and with respect to other four multiobjective‐based constraint‐handling techniques. Copyright© 2004 John Wiley & Sons, Ltd.