Explore chapters and articles related to this topic
Information Visualization
Published in Alexandru Telea, Data Visualization, 2014
As a second example, consider the information stored during the development of a software project. This consists of the actual source code files, which contain program text usually structured, from high to low level, as a set of classes, functions, and individual statements. Besides the source code, other data items can be present, such as requirements, architectural and design diagrams, development log files, and bug traces. Figure 11.1(b) sketches an example of a simple dataset containing information about a software project. This dataset contains four data elements, or tuples. Tuples p1 and p2 describe two classes and , the first written in Java and the second in C++, and having 300 and 500 lines of code, respectively. Tuples p3 and p4 describe two functions and , the first having 50 and the second 100 lines of code, respectively. The dataset additionally contains four relations, describing the fact that class contains the function , class contains the function , and that calls . These relations are indicated by the four arrows drawn in Figure 11.1(b).
Computer-Aided Control System Design: Techniques and Tools
Published in Naim A. Kheir, Systems Modeling and Computer Simulation, 2018
François E. Cellier, C. Magnus Rimvall
What about new facilities offered in future CACSD tools? We expect to see more and more flexibility with respect to the data interface. The ultimate data-driven programming is a language in which there is essentially no longer a difference between code and data. Each operation that can be performed in the language is itself expressed as an entry in a database and can thus be altered at any moment.
GAN-MP hybrid heuristic algorithm for non-convex portfolio optimization problem
Published in The Engineering Economist, 2019
Yerin Kim, Daemook Kang, Mingoo Jeon, Chungmok Lee
The computational experiments were performed on an Intel Xeon 3.5 GHz MacPro machine with 32GB memory. The proposed algorithm was implemented in Python 3.6. ILOG CPLEX 12.6 and PyTorch (http://pytorch.org/) were used as an MIP solver and a deep learning framework, respectively. To compare the performance of the proposed algorithm, the MINLP formulation obtained by Proposition 1 and 2 was solved using a nonlinear solver Couenne (https://projects.coin-or.org/Couenne/). As the machine can execute 12 threads concurrently, the algorithm employs 12 independent GANs for Steps 1 and 2 (see Figure 5). We did not employ GPU-based computation for training of GANs. The source code and data used in this paper are freely accessible on-line (https://github.com/mody3062/GAN-MP-Hybrid-Heuristic).
A systematic mapping study on enterprise architecture mining
Published in Enterprise Information Systems, 2019
Ricardo Perez-Castillo, Francisco Ruiz-Gonzalez, Marcela Genero, Mario Piattini
As a result, in order to build and maintain EA models, it is essential to determine the possible sources of knowledge. The common sources usually contain the organization’s employees, who deliver the information through experts’ interviews or serious games; documentation and manuals regarding processes, rules, responsibilities and structures in the organization, as well as technical documents concerning applications, components and IT infrastructures. Other powerful source of knowledge comes from the ISs, since these ISs provide information related to both structural and behavioural perspectives that can be extracted by analysing IS artefacts (source code, interfaces, data models, configuration files) (Sánchez, Reyes, and Villalobos 2014). ISs are extremely valuable when modelling EA since ‘they improve information access and coordination across organizational units’ and are, therefore, the most effective enablers of EA elicitation (Loucopoulos et al. 2016).
Combining data reduction, MIP solver and iterated local search for generalized assignment
Published in International Journal of Management Science and Engineering Management, 2022
Let us consider a specific MAP and its mathematical formulation (which can also be found in Soto (2011) and Soto, Rossi, and Sevaux (2012)). The considered embedded system includes memory banks, each with a limited capacity , and external memory with an assumed unlimited capacity (i.e. ). On the other hand, a computer code (for example, a C source code) involving data structures is run on the system. Data structure has size (generally in kilobytes). Accessing data structure from any memory bank takes time while it takes much more to access it from the external memory. The access time of data structure from the external memory is where is a given number. There are conflicting pairs of data structures . Two data structures are said to be conflicting if they must be accessed simultaneously. Furthermore, their conflict is said to be closed if they are allocated to the same memory bank, in which case the conflict generates an extra access time since they must be accessed sequentially. The conflict between these two data structures is open if they are allocated to two distinct memory banks. In this case, the extra access time is null.