Sequence Optimization Method Research Articles

Dimension and Stacking Sequence Optimization of Large Composite Structures Using Kriging and MOGA Method

Laminated composite structures are increasingly used for aerospace structures, such as interstage structures of launch vehicles. For an optimal design of the laminated composite structures, both dimensions and stacking sequences of the structures should be optimized simultaneously since they affect each other. The stacking sequence optimization of composite laminates is a combinatorial optimization problem with some constraints. That makes the optimization of dimensions and stacking sequences a complex mixed problem that contains both combinatorial and continuous-discrete variable optimizations. Therefore, the design of the laminated composite structures has generally been carried out using a genetic algorithm (GA). GA inherently requires many evaluations of optimization functions during the process. It must be time consuming especially for composite structural optimizations, which needs a computationally expensive analysis for the evaluation of the functions such as stress and buckling load. Authors have proposed an effective and global optimization method of the composite structures. The method utilizes a kriging method for approximating expensive functions of the optimization. Using the kriging models makes it possible not only to reduce the computational cost to evaluate the functions but also to search for the optimal design globally with using multi-objective genetic algorithm (MOGA). In the method, fractal branch and bound (FBB) method is also adopted, which is a practical and low-cost stacking sequence optimization method. In the present paper, the proposed method is applied to a complicated design problem of a large-sized composite structure. A design of a rocket structure under combined load is addressed here. The result shows that the method is effective in the large-sized optimization problems.

Extended fractal branch and bound method for optimization of multiple stacking sequences of stiffened composite panel

Stiffened composite panels are often used as structural components in aircraft in order to avoid buckling. It is well known that stacking sequence optimizations are indispensable for laminated composite structures. Stiffened composite panels usually have more than two stacking sequences because they consist of a panel skin laminate and stiffener laminates. This means that the stacking sequences need to be jointly optimized to achieve structural optimization of the stiffened composite panel. The authors have proposed a new stacking sequence optimization method, called the fractal branch and bound method, for optimizing a single laminate. In the present study, the fractal branch and bound method is extended to optimizing multiple stacking sequences. The extended method is applied for obtaining two optimal stacking sequences for the maximization of the buckling load of a hat-stiffened composite panel. The improved method successfully provides two optimal stacking sequences determined in a short period of time.

Stacking Sequence Optimizations of Composite Structures Using Fractal Branch and Bound Method. Optimization Problem Includes In-Plane and Out-of-Plane Lamination Parameters.

In the present study, a mechanism of fractal image of laminates on lamination parameters is discussed in detail, and a stacking sequence optimization method using the mechanism is proposed for complicated structures that include both of in-plane and out-of-plane lamination parameters. The new method employs branch and bound method for the optimizations of stacking sequences. For the estimation of the fractal branch of stacking sequences, the new method requires approximation of the objective function of the optimizations with quadratic polynomials using both in-plane and out-of-plane lamination parameters. The new method is applied to a stacking sequence optimization problem of a maximization of buckling load of a hat type stringer structure. The method gives successfully optimal stacking sequences in a short time.

Monte Carlo procedure for protein design

A new method for sequence optimization in protein models is presented. The approach, which has inherited its basic philosophy from recent work by Deutsch and Kurosky [Phys. Rev. Lett. 76, 323 (1996)] by maximizing conditional probabilities rather than minimizing energy functions, is based upon a novel and very efficient multisequence Monte Carlo scheme. By construction, the method ensures that the designed sequences represent good folders thermodynamically. A bootstrap procedure for the sequence space search is devised making very large chains feasible. The algorithm is successfully explored on the two-dimensional HP model with chain lengths N=16, 18 and 32.

Development of simple fitness landscapes for peptides by artificial neural filter systems

The applicability of artificial neural filter systems as fitness functions for sequence-oriented peptide design was evaluated. Two example applications were selected: classification of dipeptides according to their hydrophobicity and classification of proteolytic cleavage-sites of protein precursor sequences according to their mean hydrophobicities and mean side-chain volumes. The cleavage-sites covered 12 residues. In the dipeptide experiments the objective was to separate a selected set of molecules from all other possible dipeptide sequences. Perceptrons, feedforward networks with one hidden layer, and a hybrid network were applied. The filters were trained by a (1, lambda) evolution strategy. Two types of network units employing either a sigmoidal or a unimodal transfer function were used in the feedforward filters, and their influence on classification was investigated. The two-layer hybrid network employed gaussian activation functions. To analyze classification of the different filter systems, their output was plotted in the two-dimensional sequence space. The diagrams were interpreted as fitness landscapes qualifying the markedness of a characteristic peptide feature which can be used as a guide through sequence space for rational peptide design. It is demonstrated that the applicability of neural filter systems as a heuristic method for sequence optimization depends on both the appropriate network architecture and selection of representative sequence data. The networks with unimodal activation functions and the hybrid networks both led to a number of local optima. However, the hybrid networks produced the best prediction results. In contrast, the filters with sigmoidal activation produced good reclassification results leading to fitness landscapes lacking unreasonable local optima. Similar results were obtained for classification of both dipeptides and cleavage-site sequences.

Topics on speech waveform coding activities in Japan

Current activities on speech waveform coding technique developments in Japan are reviewed. The following topics of the activities on ADPCM and ADM are discussed. (1) An ADPCM code sequence optimization method by Fushikida for an ADPCM with an adaptive quantization and fixed predictor. The method uses a technique reminiscent of dynamic programming and optimizes code sequence. It is shown that about 3 dB SNR improvement is obtainable for speech at 32 kbps (4 bits/sample). (2) Another ADPCM technique by Araseki and Ochiai. This technique is suitable for use at the bit rate less than 32 kbps. Two adaptive predictors connected in a cascaded form are used. One of the predictors has an ordinary small number of prediction taps. However, the other predictor possesses a considerably larger number of predictor taps and is used to predict the speech waveform which is one pitch period apart. (3) Performance improvement techniques and LSI implementation by Tomozawa and Niwa for a discretely variable slope 32 kbps ADM (DVSD). Yatsuzuka’s residual encoder and Hosokawa and Yamashita’s adaptive predictive DM are also discussed briefly.

Topics on speech waveform coding activities in Japan

Topics on ADPCM and ADM technique developments in Japan are described. First, an ADPCM code sequence optimization method by K. Fushikida is discussed for an ADPCM with an adaptive quantization and fixed predictor. The method uses a technique reminiscent of dynamic programming and optimizes code sequence. It is shown that about 2‐dB SNR improvement is obtainable for speech at 32 kbps (4 bits/sample). Second, another ADPCM technique by T. Araseki and K. Ochiai is described. This technique is suitable for use at the bit rate less than 32 kbps. Two adaptive predictors connected in a cascaded form are used. One of the predictors has an ordinary small number of prediction taps. However, the other predictor possesses a considerably large number of predictor taps which are used to predict speech waveform in one pitch period apart. A backward adaptive prediction technique with a gradient method is used for each of the two adaptive predictors. Finally, performance improvement techniques and LSI implementation by A. Tomozawa and K. Niwa are described for a discretely variable slope 32 kbps ADM (DVSD).