China
Africa
Explore chapters and articles related to this topic
Simulation and Example
Published in Jiguo Yu, Xiuzhen Cheng, Honglu Jiang, Dongxiao Yu, Hierarchical Topology Control for Wireless Networks, 2018
Jiguo Yu, Xiuzhen Cheng, Honglu Jiang, Dongxiao Yu
For the development of the C++ program, after analyzing and abstracting the problem, the problem model is built. Then the model is transformed into a finite sequence of steps. The statement sequence of C++ is composed according to its grammar, that is, the C++ source code. The source code from the keyboard input to the computer is saved as a file of .cpp format. This process is known as editing. Then the source file is compiled into the .OBJ file and then connected to the .EXE executable file. The processed results can be obtained by running the executable file (Figure 5.28).
A crystallization case study toward optimization of expensive to evaluate mathematical models using Bayesian approach
Published in Materials and Manufacturing Processes, 2023
Aswitha Tadepalli, Keerthi NagaSree Pujari, Kishalay Mitra
To integrate the PB model (Section 3.1) with MOBO in Python, it was converted into an executable file (.exe). This executable file was called iteratively to generate the respective outputs required to train the MOBO algorithm’s GPR surrogate during the optimization process. Initially, using the Sobol sampling technique, for 10 decision variables, 2(10 + 1) = 22 [as commonly suggested as 2(# of decision variables + 1)] samples were generated. Since MOBO is an iterative process, a maximum number of 150 iterations was set as the budget for optimization exercise, also used as the termination criteria. Since the objectives were evaluated using a high fidelity PB model, there will be a minimum scope for noisy observations as opposed to cases where such objective values are achieved through experiments. Thus, for the PB model, the MOBO algorithm was run for 150 iterations with q-EHVI without noisy observations as the acquisition function. To ensure the reproducibility of the results, a seed was set in the BoTorch® package to give the same set of initial points each time. Figure 2 shows the hypervolume improvement plot in the MOBO algorithm and the Pareto fronts obtained at different iterations are presented in Fig. 3.
Trigger motion and interface optimization of an eye-controlled human-computer interaction system based on voluntary eye blinks
Published in Human–Computer Interaction, 2023
Guo-Rui Ma, Jia-Xin He, Chun-Hsien Chen, Ya-Feng Niu, Lan Zhang, Tian-Yu Zhou
The experimental stimulus was presented by a Unity3D-based EXE executable. There were 3 stages in each trial. At stage 1, a cross mark in the center of the screen lasting for 1000 ms was presented, to guide the subject’s line of sight back to the central area of the screen. At Stage 2, the white circular IO appeared on the screen and the subject triggered the IO with the required motion combination according to different IO designs within 3000 ms time pressure. At stage 3, after triggering, the screen immediately jumped to the empty screen page for 1000 ms and the background color remained unchanged to eliminate the visual persistence. If the subject failed to trigger the IO within 3000 ms, a failure count would be added to the counter. The operation process of a single trial is shown in Figure 3.
Adaptive data-driven optimization of chiller loading with domain knowledge
Published in Science and Technology for the Built Environment, 2021
Wantao Shi, Xinqiao Jin, Jiangqing Wang, Zhimin Du
Figure 7 shows the online strategies implementation platform. The proposed strategy and Strategy 3 are programmed in Python 3.6, and they are compiled as EXE (executable program). The EXE is installed on the supervisory computer. The data acquisition and control modules are installed on the local computer. The optimization strategies and the modules exchange data through a database, and the flowchart of data exchange is described below. The optimization strategies read the current operation condition of the multi-chiller system from the database and calculate the optimal setpoints of chilled water supply temperature and chillers’ on/off states under the current operation condition. Then the setpoints are written into the database. On the other hand, the data acquisition and control modules write the current operation condition collected by the sensors into the database and read the latest setpoints from the database every 1 minute. Then the modules control the chillers’ on-off and the setpoints adaptively.