An Adaptive Coevolution Method for Efficient Robust Optimization Under Interval Uncertainty

This paper presents a new approach aims to solve robust multidisciplinary design optimization MDO problem called Improved Multi-objective Robust Collaborative Optimization . This method combines the Multi-objective Robust Collaborative Optimization method, the Worst Possible Point constraint cuts and the Genetic algorithm NSGA-II type as an optimizer to solve the robust optimization problem of complex structure named Y-stiffened panel under interval uncertainty . The proposed approach hierarchically decomposes the optimization problem into a structure level considered as an upper level in the Y-stiffened panel and a second level considered as a lower level of the studied panel. A robust multi-objective optimization problem intended to optimize the eigenfrequency, the global mass and the displacement at a fixed point of the Y-stiffened panel at the first level and each structure’s robust optimization problem allows optimizing its eigenfrequency and mass limited by their local constraint functions at the second one. Tor demonstrate our method, an engineering example of Y-stiffened panel is treated. A good performance of proposed method is proved by a comparison between obtained results and Non-Distributed Multi-objective Robust Optimization .

Surrogate duality for robust optimization

A number of multi-objective evolutionary algorithms have been proposed in recent years and many of them have been used to solve engineering design optimization problems. However, designs need to be robust for real-life implementation, i.e. performance should not degrade substantially under expected variations in the variable values or operating conditions. Solutions of constrained robust design optimization problems should not be too close to the constraint boundaries so that they remain feasible under expected variations. A robust design optimization problem is far more computationally expensive than a design optimization problem as neighbourhood assessments of every solution are required to compute the performance variance and to ensure neighbourhood feasibility. A framework for robust design optimization using a surrogate model for neighbourhood assessments is introduced in this article. The robust design optimization problem is modelled as a multi-objective optimization problem with the aim of simultaneously maximizing performance and minimizing performance variance. A modified constraint-handling scheme is implemented to deal with neighbourhood feasibility. A radial basis function (RBF) network is used as a surrogate model and the accuracy of this model is maintained via periodic retraining. In addition to using surrogates to reduce computational time, the algorithm has been implemented on multiple processors using a master–slave topology. The preliminary results of two constrained robust design optimization problems indicate that substantial savings in the actual number of function evaluations are possible while maintaining an acceptable level of solution quality.

In this paper, we design a Branch and Bound algorithm based on interval arithmetic to address nonconvex robust optimization problems. This algorithm provides the exact global solution of such difficult problems arising in many real life applications. A code was developed in MatLab and was used to solve some robust nonconvex problems with few variables. This first numerical study shows the interest of this approach providing the global solution of such difficult robust nonconvex optimization problems.

Fast robust optimization of ORC based on an artificial neural network for waste heat recovery

Mars entry trajectory planning using robust optimization and uncertainty quantification

In Chapter 4, several polynomially solvable cases of robust discrete optimization problems are discussed in detail. However, by results presented in Chapter 3 we also know that most robust discrete optimization problems belong to the NP-hard class. In this chapter, we present our approach for solving these difficult robust discrete optimization problems. We are in this chapter restricting our attention to robust discrete optimization problems with equivalent single scenario problems that can be efficiently solved with a polynomial or pseudo-polynomial procedure. The solution procedures are based on branch-and-bound with both upper and lower bounds generated by surrogate relaxation. To be exact, the upper bound (for a maximization problem) is obtained from surrogate relaxation and the lower bound is obtained as a by-product via a heuristic based on the surrogate relaxation result. In the case when input data satisfies bounded percentage deviation condition (to be defined in Section 5.2), the heuristic is shown to provide a constant approximation. Computational results in Section 5.3 demonstrate the effectiveness of the bounds and the solution procedure.

Robust optimization techniques attempt to find a solution that is both optimum and relatively insensitive to input uncertainty. In general, these techniques are computationally more expensive than their deterministic counterparts. In this article two new robust optimization methods are presented. The first method is called gradient-assisted robust optimization (GARO). In GARO, a robust optimization problem is first converted to a deterministic one by using a gradient-based approximation technique. After solving this deterministic problem, the solution robustness and the accuracy of the approximation are checked. If the accuracy meets a threshold, a robust optimum solution is found; otherwise, the approximation is adaptively modified until the threshold is met and a solution, if it exists, is obtained. The second method is a faster version of GARO called quasi-concave gradient-assisted robust optimization (QC-GARO). QC-GARO is for problems with quasi-concave objective and constraint functions. The difference between GARO and QC-GARO is in the way that they check the approximation accuracy. Both GARO and QC-GARO methods are applied to a set of six engineering design test problems and the results are compared with a few related previous methods. It was found that, compared to the methods considered, GARO could solve all test problems but with a higher computational effort compared to QC-GARO. However, QC-GARO was computationally much faster when it was able to solve the problems.

In this paper, we define a distributionally robust k-sum optimization problem as the problem of finding a solution that minimizes the worst-case expected sum of up to the k largest costs of the elements in the solution. The costs are random with a joint probability distribution that is not completely specified but rather assumed to be known to lie in a set of probability distributions. For k=1, this reduces to a distributionally robust bottleneck optimization problem while for k=n, this reduces to distributionally robust minimum sum optimization problem. Our main result is that for a Fréchet class of discrete marginal distributions with finite support, the distributionally robust k-sum combinatorial optimization problem is solvable in polynomial time if the deterministic minimum sum problem is solvable in polynomial time. This extends the result of Punnen and Aneja [On k-sum optimization, Oper. Res. Lett. 18(5)(1996), pp. 233–236] from the deterministic to the robust case. We show that this choice of the set of distributions helps preserve the submodularity of the k-sum objective function which is an useful structural property for optimization problems.

A robust optimum solution is defined as one that is optimum and maintains its feasibility regardless of the values taken by uncertain parameters within a bounded interval. Solving for a robust optimum solution, however, can be computationally costly. To reduce the computational cost, a new robust optimization method, surrogate feasibility testing–cutting robust optimization (SFTC-RO), is proposed that solves two optimization subproblems. In the first subproblem, a scenario robust optimization problem is solved using a local optimization technique. In the second subproblem, a feasibility check and constraint cut of the feasible domain are performed via operations on a surrogate model of the constraints. The two subproblems are iteratively solved until convergence at a robust optimum solution. Results from numerical and engineering examples show that on average the proposed method can obtain robust optimal solutions at a lower computational cost and with better scalability than an existing method from the literature.

Proton therapy is preferred for its dose conformality to spare normal tissues and organs-at-risk (OAR) via Bragg peaks with negligible exit dose. However, proton dose conformality can be further optimized: (1) the spot placement is based on the structured (e.g., Cartesian) grid, which may not offer conformal shaping to complex tumor targets; (2) the spot sampling pattern is uniform, which may be insufficient at the tumor boundary to provide the sharp dose falloff, and at the same time may be redundant at the tumor interior to provide the uniform dose coverage, for example, due to multiple Coulomb scattering (MCS); and (3) the lateral spot penumbra increases with respect to the depth due to MCS, which blurs the lateral dose falloff. On the other hand, while (1) the deliverable spots are subject to the minimum-monitor-unit (MMU) constraint, and (2) the dose rate is proportional to the MMU threshold, the current spot sampling method is sensitive to the MMU threshold and can fail to provide satisfactory plan quality for a large MMU threshold (i.e., high-dose-rate delivery). This work will develop a novel Triangular-mEsh-based Adaptive and Multiscale (TEAM) proton spot generation method to address these issues for optimizing proton dose conformality and plan delivery efficiency. Compared to the standard clinically-used spot placement method, three key elements of TEAM are as follows: (1) a triangular mesh instead of a structured grid: the triangular mesh is geometrically more conformal to complex target shapes and therefore more efficient and accurate for dose shaping inside and around the target; (2) adaptive sampling instead of uniform sampling: the adaptive sampling consists of relatively dense sampling at the tumor boundary to create the sharp dose falloff, which is more accurate, and coarse sampling at the tumor interior to uniformly cover the target, which is more efficient; and (3) depth-dependent sampling instead of depth-independent sampling: the depth-dependent sampling is used to compensate for MCS, that is, with increasingly dense sampling at the tumor boundary to improve dose shaping accuracy, and increasingly coarse sampling at the tumor interior to improve dose shaping efficiency, as the depth increases. In the TEAM method the spot locations are generated for each energy layer and layer-by-layer in the multiscale fashion; and then the spot weights are derived by solving the IMPT problem of dose-volume planning objectives, MMU constraints, and robustness optimization with respect to range and setup uncertainties. Compared to the standard clinically-used spot placement method UNIFORM, TEAM achieved (1) better plan quality using <60% number of spots of UNIFORM; (2) better robustness to the number of spots; (3) better robustness to a large MMU threshold. Furthermore, TEAM provided better plan quality with fewer spots than other adaptive methods (Cartesian-grid or triangular-mesh). A novel triangular-mesh-based proton spot placement method called TEAM is proposed, and it is demonstrated to improve plan quality, robustness to the number of spots, and robustness to the MMU threshold, compared to the clinically-used spot placement method and other adaptive methods.

Adaptive and robust radiation therapy optimization for lung cancer

We propose a dynamical system-based approach for solving robust optimization problems. The well-known continuous-time dynamical system for solving deterministic optimization problems arises in the form of primal-dual gradient dynamics where the vector field is derived as the gradient of the Lagrangian. The new continuous-time dynamical system we introduce for solving robust optimization problems differs from the primal-dual dynamics in the sense that the vector field is not derived as the gradient of the Lagrangian function. We call this new dynamical system as saddle point dynamics. In the saddle point dynamics, the uncertain variable arises as a dynamical state. For a general class of robust optimization problem, where the cost function is convex in decision variable and concave in uncertain variable, we show that the robust optimal solution can be recovered as a globally asymptotically stable equilibrium point of the saddle point dynamical system. Simulation results are presented to demonstrate the capability of this new dynamical system to solve various robust optimization problems. We also compare our proposed approach with existing methods based on robust counterpart and scenario-based random sampling.

Quadratically adjustable robust linear optimization with inexact data via generalized S-lemma: Exact second-order cone program reformulations

Chapter 4 - Data-driven robust stochastic optimization for power systems operations