A Deep Learning Approach for Maximum Activity Links in D2D Communications

Mobile cellular communications are experiencing an exponential growth in traffic load on Long Term Evolution (LTE) eNode B (eNB) components. Such load can be significantly contained by directly sharing content among nearby users through device-to-device (D2D) communications, so that repeated downloads of the same data can be avoided as much as possible. Accordingly, for the purpose of improving the efficiency of content sharing and decreasing the load on the eNB, it is important to maximize the number of simultaneous D2D transmissions. Specially, maximizing the number of D2D links can not only improve spectrum and energy efficiency but can also reduce transmission delay. However, enabling maximum D2D links in a cellular network poses two major challenges. First, the interference between the D2D and cellular communications could critically affect their performance. Second, the minimum quality of service (QoS) requirement of cellular and D2D communication must be guaranteed. Therefore, a selection of active links is critical to gain the maximum number of D2D links. This can be formulated as a classical integer linear programming problem (link scheduling) that is known to be NP-hard. This paper proposes to obtain a set of network features via deep learning for solving this challenging problem. The idea is to optimize the D2D link schedule problem with a deep neural network (DNN). This makes a significant time reduction for delay-sensitive operations, since the computational overhead is mainly spent in the training process of the model. The simulation performed on a randomly generated link schedule problem showed that our algorithm is capable of finding satisfactory D2D link scheduling solutions by reducing computation time up to 90% without significantly affecting their accuracy.

Assigning Examination Invigilator's Schedule Problem: An Approach of Programming Techniques Using Simple Sequential Assignment

Preparation of examination invigilation schedule is a tedious and challenging tasks. It is impossible to satisfy all invigilators or proctors with the invigilation schedules prepared. The examination committees have to analyze and thoroughly checks all constraints submitted by the invigilators. Formerly, the process of assignments and deciding the best slot for invigilation will usually take a couple of weeks and all these processes are done manually. This paper proposes a programming technique using simple sequential approach to assign the invigilators on the proper examination slot without any conflicts or clashes aligned with the list of constraints and parameters. This simple tools or systems has been developed to help the examination committee to reduce the time taken for invigilation schedule preparation, avoid erroneous of incorrect assignment of invigilators and increase the satisfaction amongst invigilators with the invigilation schedule assigned. However, this tool is only a supplementary support in invigilation assignment process and the manual changes by considering human touch factors are still considered to produce friendly and empathy worthy invigilation schedules.

A Comparison of Genetic Algorithm, Particle Swarm Optimization and Simulated Annealing in Real-Time Task Scheduling on CMP

Real-time task schedule problem in Chip-Multiprocessor (CMP) receives wide attention in recent years. It is partly because the increasing demand for CMP solutions call for better schedule algorithm to exploit the full potential of hardware, and partly because of the complexity of schedule problem, which itself is an NP-hard problem. To address this task schedule problem, various of heuristics have been studied, among which, Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Simulated Annealing (SA) are the most popular ones. In this paper, we implement these 3 schedule heuristics, and compare their performance under the context of real-time tasks scheduling on CMP. According to the results of our intensive simulations, PSO has the best fitness optimization of these 3 algorithms, and SA is the most efficient algorithm.