In the literature on generalized pattern search algorithms, convergence to a stationary point of a once continuously differentiable cost function is established under the assumption that the cost function can be evaluated exactly. However, there is a large class of engineering problems where the numerical evaluation of the cost function involves the solution of systems of differential algebraic equations. Since the termination criteria of the numerical solvers often depend on the design parameters, computer code for solving these systems usually defines a numerical approximation to the cost function that is discontinuous with respect to the design parameters. Standard generalized pattern search algorithms have been applied heuristically to such problems, but no convergence properties have been stated. In this paper we extend a class of generalized pattern search algorithms to include a subprocedure that adaptively controls the precision of the approximating cost functions. The numerical approximations to the cost function need not define a continuous function. Our algorithms can be used for solving linearly constrained problems with cost functions that are at least locally Lipschitz continuous. Assuming that the cost function is smooth, we prove that our algorithms converge to a stationary point. Under the weaker assumption that the cost function is only locally Lipschitz continuous, we show that our algorithms converge to points at which the Clarke generalized directional derivatives are nonnegative in predefined directions. An important feature of our adaptive precision scheme is the use of coarse approximations in the early iterations, with the approximation precision controlled by a test. We show by numerical experiments that such an approach leads to substantial time savings in minimizing computationally expensive functions.

VL - 16 IS - 3 ER - TY - JOUR T1 - Building design optimization using a convergent pattern search algorithm with adaptive precision simulations JF - Energy and Buildings Y1 - 2005 A1 - Michael Wetter A1 - Elijah Polak AB -We propose a simulation–precision control algorithm that can be used with a family of derivative free optimization algorithms to solve optimization problems in which the cost function is defined through the solutions of a coupled system of differential algebraic equations (DAEs). Our optimization algorithms use coarse precision approximations to the solutions of the DAE system in the early iterations and progressively increase the precision as the optimization approaches a solution. Such schemes often yield a significant reduction in computation time. We assume that the cost function is smooth but that it can only be approximated numerically by approximating cost functions that are discontinuous in the design parameters. We show that this situation is typical for many building energy optimization problems.We present a new building energy and daylighting simulation program, which constructs approximations to the cost function that converge uniformly on bounded sets to a smooth function as precision is increased.We prove that for our simulation program, our optimization algorithms construct sequences of iterates with stationary accumulation points. We present numerical experiments in which we minimize the annual energy consumption of an office building for lighting, cooling and heating. In these examples, our precision control algorithm reduces the computation time up to a factor of four.

VL - 37 IS - 6 U2 - LBNL-57341 ER - TY - JOUR T1 - Building Design Optimization Using a Convergent Pattern Search Algorithm with Adaptive Precision Simulations JF - Energy and Buildings Y1 - 2004 A1 - Michael Wetter A1 - Elijah Polak VL - 37 ER - TY - RPRT T1 - BuildOpt 1.0.1 validation Y1 - 2004 A1 - Michael Wetter A1 - Elijah Polak A1 - Van P. Carey U2 - LBNL-54658 ER - TY - JOUR T1 - A convergent optimization method using pattern search algorithms with adaptive precision simulation JF - Building Services Engineering Research and Technology Y1 - 2004 A1 - Michael Wetter A1 - Elijah Polak AB -Thermal building simulation programs, such as EnergyPlus, compute numerical approximations to solutions of systems of differential algebraic equations. We show that the exact solutions of these systems are usually smooth in the building design parameters, but that the numerical approximations are usually discontinuous due to adaptive solvers and finite precision computations. If such approximate solutions are used in conjunction with optimization algorithms that depend on smoothness of the cost function, one needs to compute high precision solutions, which can be prohibitively expensive if used for all iterations. For such situations, we have developed an adaptive simulation–precision control algorithm that can be used in conjunction with a family of derivative free optimization algorithms. We present the main ingredients of the composite algorithms, we prove that the resulting composite algorithms construct sequences with stationary accumulation points, and we show by numerical experiments that using coarse approximations in the early iterations can significantly reduce computation time.

VL - 25 IS - 4 ER - TY - Generic T1 - A convergent optimization method using pattern search algorithms with adaptive precision simulation T2 - Proceedings of the 8th IBPSA Conference Y1 - 2003 A1 - Michael Wetter A1 - Elijah Polak ED - Godfried Augenbroe ED - Jan Hensen KW - coordinate search KW - direct search KW - genetic algorithm KW - hooke–jeeves KW - optimization KW - particle swarm optimization AB -In solving optimization problems for building design and control, the cost function is often evaluated using a detailed building simulation program. These programs contain code features that cause the cost function to be discontinuous. Optimization algorithms that require smoothness can fail on such problems. Evaluating the cost function is often so time-consuming that stochastic optimization algorithms are run using only a few simulations, which decreases the probability of getting close to a minimum. To show how applicable direct search, stochastic, and gradient-based optimization algorithms are for solving such optimization problems, we compare the performance of these algorithms in minimizing cost functions with different smoothness. We also explain what causes the large discontinuities in the cost functions.

JF - Proceedings of the 8th IBPSA Conference CY - Eindhoven, Netherlands VL - III ER -