China
Africa
Explore chapters and articles related to this topic
Principles of Software Design and Development
Published in Sharon Yull, BTEC National for IT Practitioners: Core Units, 2009
Debug tools are used to detect errors within a programme and track the data flow to address and correct logic or design errors. Debug tools can also be used to monitor the value of variables at certain stages within the programme using ‘break points’.
BFS: A blockchain-based financing scheme for logistics company in supply chain finance
Published in Connection Science, 2022
Jia Fu, Bangcan Cao, Xiaoliang Wang, Pengjie Zeng, Wei Liang, Yuzhen Liu
As can be seen in Figure 10, at a block size of 1024KB, the throughput steadily increases when the QPS is between 0 and 200, independent of the log level. After that, the rising trend of throughput becomes slow and stabilises. As can be seen, the higher the log level, the higher the throughput of the system. When the logging level is set to DEBUG, the stable value of throughput is very different from the other three levels. The reason for this is that the lower the log level in Fabric, the more detailed the log output will be, and the DEBUG level is the lowest log level, making it easier for system administrators to debug problems after they have occurred. However, too much logging information output can reduce the throughput performance of the system. It can be seen that BFS has better transfer performance under the same log level condition. This is because BFS can find the demanded transaction information from the data storage module faster in a stable topology and can optimise many repetitive transaction logics in the related operations in the supply chain finance scenario, which can significantly improve the transaction efficiency and enhance the financing efficiency of logistics companies.
Robot behavior debugger for non-expert users in convenience stores using behavior trees
Published in Advanced Robotics, 2022
Pattaraporn Tulathum, Bunyapon Usawalertkamol, Gustavo Alfonso Garcia Ricardez, Jun Takamatsu, Tsukasa Ogasawara, Kenichi Matsumoto
This section presents the details of our proposed Behavior Tree-based system. Figure 2 shows the proposed system components including the robot behavior creator and the robot behavior debugger. To use the robot behavior creator, expert users need to prepare sets of human actions and robot reactions. Then, non-expert users can create a robot behavior program by making pairs of human actions and robot reactions from the prepared sets. After that, non-expert users can test their created programs via the robot simulator to see how a robot responds to a customer in a simulated convenience store environment. When the non-expert users observe that there is an unexpected result, they can use the proposed debugger to debug the program via four components including setting breakpoints, logging node status, monitoring node execution, and robot status variables. Examples of our system usage are shown in the following link: https://youtube.com/playlist?list=PLCMeFFUGH8tqqzykNbLwf6uudgF1yLLsN.
Implementation issues in optimization algorithms: do they matter?
Published in Journal of Experimental & Theoretical Artificial Intelligence, 2019
Thomas Weise, Yuezhong Wu, Weichen Liu, Raymond Chiong
Based on the experience of implementing LK Heuristics in the first case study, we were able to avoid unnecessary effort when implementing the ECMs: We directly designed efficient data structures, and accepted the higher learning curve compared to implementing more complex algorithm versions step-by-step. This saved much time and reduced the number of bugs in the code. It should be noted that this experience does not fit very well to common software development procedures, where a more iterative development process is often preferred, similar to the first case study. Different from the situation in common software projects, however, our requirements are straightforward, clear, and do not change. As in common software projects, JUnit tests were used to debug every modification of the implementations and greatly helped us in producing the correct algorithms.