Search

In this paper, a teaching-learning-based optimization algorithm (TLBO) is proposed to solve the multi-skill resource constrained project scheduling problem (MS-RCPSP) with makespan minimization criterion. A task-resource list-based encoding scheme is presented by combining the task list and the resource list, and a left-shift decoding scheme is developed to generate feasible schedules. To achieve satisfactory performances, the balance between global exploration and local exploitation is stressed in designing the TLBO algorithm. At the initialization stage, a balanced resource rule is proposed to generate the initial resource lists, and multiple task list rules are adopted in a hybrid way to initialize the task lists. At the teacher phase and the student phase, the two-point crossover and the resource-based local search are utilized to generate the promising task-resource lists. Moreover, a reinforcement phase is incorporated into the original TLBO with both the permutation-based and the resource-based local search strategies as an additional phase to enhance the local intensification capability. To investigate the influence of parameter setting on the TLBO, numerical tests based on Taguchi method of design of experiment are carried out. In addition, the effectiveness of the proposed balanced resource rule is shown by statistical comparisons with the random resource rule. Computational comparisons between TLBO and the existing algorithm also demonstrate the effectiveness and efficiency of the proposed TLBO in solving the MS-RCPSP. [ABSTRACT FROM AUTHOR]

Wireless sensor networks (WSNs) have been applied in a variety of application areas. Most WSN systems, once deployed, are intended to operate unattended for a long period. During the lifetime, it is necessary to fix bugs, reconfigure system parameters, and upgrade the software in order to achieve reliable system performance. However, manually collecting all nodes back and reconfiguring through serial connections with computer is infeasible since it is labor-intensive and inconvenient due to the harsh deploying environments. Hence, data dissemination over multi-hop is desired to facilitate such tasks. This survey discusses the requirements and challenges of data dissemination in WSNs, reviews existing work, introduces some relevant techniques, presents the metrics of the performance and comparisons of the state-of-the-art work, and finally suggests the possible future directions in data dissemination studies. This survey elaborates and compares existing approaches of two categories: structure-less schemes and structure-based schemes, classified by whether or not the network structure information is used during the disseminating process. In existing literatures, different categories have definite boundary and limited analysis on the trade-off between different categories. Besides, there is no survey that discusses the emerging techniques such as Constructive Interference (CI) while these techniques have the chance to change the framework of data dissemination. In a word, even though many efforts have been made, data dissemination in WSNs still needs some more work to embrace the new techniques and improve the efficiency and practicability further. [ABSTRACT FROM AUTHOR]