Improving the accuracy of forecasting is crucial but complex in clothing industry, especially for new products, with lack historical data and a wide range factors affecting demand. Previous studies more concentrate on sales rather than demand forecasting, variables remained to be optimized. In this study, two-stage intelligent retail system designed products. first stage, estimated original considering stock-out. The adaptive neuro fuzzy inference (ANFIS) introduced into second stage...

Machine learning is applied to investigate the phase transition of two-dimensional complex plasmas. The Langevin dynamics simulation employed prepare particle suspensions in various thermodynamic states. Based on resulted positions two extreme conditions, bitmap images are synthesized and imported a convolutional neural network (ConvNet) as training sample. As result, diagram obtained. This trained ConvNet model can be directly sequence recorded using video microscopy experiments study melting.

Laboratory realizations of two-dimensional (2D) plasma crystals typically involve monodisperse microparticles confined into horizontal monolayers in radio-frequency (rf) sheaths. This gives rise to the so-called wakes beneath microparticles. The presence renders interactions such systems nonreciprocal, a fact that can lead quite different behavior from one expected for their reciprocal counterparts. Here we examine buckling hexagonal 2D crystal, occurring as confinement strength is...

A binary complex plasma consists of two different types dust particles in an ionized gas. Due to the spinodal decomposition and force imbalance, masses diameters are typically phase separated, resulting interface. Both external excitation internal instability may cause interface move with time. Support vector machine (SVM) is a supervised learning method that can be very effective for multi-class classification. We applied SVM classification based on image brightness locate plasma. Taking...

Domains of the square lattice have been observed in a quasi-two-dimensional binary complex plasma. The longitudinal and transverse modes wave spectra were measured. To compare with experiment, Langevin dynamics simulations mixture carried out, where non-reciprocal interactions between different species modeled point-wake Yukawa potential. A strong dependence on relative magnitude charge is revealed.

The propagation of a dissipative soliton was experimentally studied in two-dimensional binary complex plasma. crystallization suppressed the center particle suspension where two types particles were mixed. motions individual recorded using video microscopy, and macroscopic properties solitons measured amorphous mixture plasma crystal periphery. Although overall shape parameters propagating crystalline regions quite similar, their velocity structures at small scales as well distributions...

Binary complex plasmas consist of microparticles two different species and can form two-dimensional square lattices under certain conditions. The dispersion relations the lattice waves are derived for longitudinal transverse in-plane modes, assuming that out-of-plane mode is suppressed by strong vertical confinement. results compared with spectra obtained in Langevin dynamics simulations. Furthermore, we investigate dependence relation on charge ratio mass particle species.

The penetration of a supersonic particle at the interface was studied in binary complex plasma. Inspired by experiments performed PK-3 Plus Laboratory on board International Space Station, Langevin dynamics simulations were carried out. evolution Mach cone observed, where kink lateral wake front observed interface. By comparing axial and radial velocity, we show that solitary wave is non-linear. dependence background vicinity direction reveals disparity mobility may be cause various effects.

Complex plasmas consist of ionized gas and charged solid microparticles, representing the plasma state soft matter. We apply machine learning methods to investigate a melting transition in two-dimensional complex plasma. A convolutional neural network is constructed trained with numerical simulation. The hexatic phase successfully identified evolution topological defects studied during both simulations experiments.

Machine learning methods have been widely used in the investigations of complex plasmas. In this paper, we demonstrate that unsupervised convolutional neural network can be applied to obtain melting line two-dimensional plasmas based on Langevin dynamics simulation results. The training samples do not need labeled. resulting coincides with those obtained by analysis hexatic order parameter and supervised machine method.

We demonstrate a direct-detection phase retrieval receiver based on silicon photonics. The implements strong dispersion and delay lines compact chip. retrieve the full field of 30-GBd QPSK signal without carrier or local oscillator.

A complex plasma is a weakly ionized gas containing electrons, ions, neutral atoms and small macroscopic particles. Such system allows experimental studies of various physical processes occurring in liquids solids at the kinetic level. In this talk, we present experiments simulations on few interfacial phenomena phase-separated binary plasma.