Teaching-learning Algorithm Research Articles

Energy efficient task distribution with efficient deployment and scheduling of sensor nodes in WSN

Lifetime of the network plays a vital role in working wireless sensor networks. Network lifetime can be improved by the efficient use of present sensor nodes in wireless sensor nodes. In this paper, our objective is to deploy and schedule the sensor nodes in such a manner that improves the lifetime and provides a way to distribute the tasks among the sensor nodes that completes the task on the basis of priority and accuracy. Thus, the overall objective of this paper is to calculate the optimised location of sensor nodes contain a fix sensing range and schedule the sensor nodes that maximises the lifetime. The upper bound of network lifetime can be pre-calculated mathematically, we are applying this knowledge to find the locations of sensor nodes that improves the lifetime. Further, the nodes are scheduled and task is distributed among the sensor nodes that achieve this upper bound. In this paper, we are applying teaching learning algorithm to find the deployment location and schedule followed by an algorithm for task distribution. The comparative study shows that our method outperforms as compared to naive methods.

A novel chaotic teaching–learning-based optimization algorithm for multi-machine power system stabilizers design problem

This paper proposes an efficient optimization algorithm named chaotic teaching–learning algorithm (CTLA), to solve multimachine power system stabilizers design problem. The original teaching learning algorithm as competitive to other optimization algorithms, used two phases to proceed to the global optimal solution: ‘teacher phase’ and ‘learner phase’. However, during the second phase an adequate interaction between the teacher and the learners in entire search space are not guaranteed and the algorithm may be trapped in local optima. Thus, in the proposed CTLA a new phase named “chaotic phase” is added in order to overcome this drawback. The performance of the CTLA is investigated by using a set of benchmark functions. To demonstrate the effectiveness of the proposed algorithm in power systems, the conventional lead-lag power system stabilizers (PSSs) are tuned for: three machines nine bus system (WSCC) and the ten machine thirty-nine bus New England power systems. The performance of the proposed CTLA-based PSS (CTLAPSSs) under different loading conditions and disturbances is investigated through eigen-value analysis, non-linear time domain-simulations and some performance indices.

Reliable/cost‐effective maintenance schedules for a composite power system using fuzzy supported teaching learning algorithm

With the increasing electricity power consumption and time-varying load demands, maintaining system reliability and economic operation has become a challenging task for power system operators during the scheduled outages. Nevertheless, very few publications in the current literature tackle this maintenance schedule problem by means of evolutionary algorithms, and, when they do, they focus only on generators. In the proposed model, the authors have considered coordinated maintenance scheduling (CMS) with cost minimisation and reliability maximisation as a multi-objective criterion, and fuzzy-based teaching learning-based optimisation is used as an optimisation tool for solving the formulated maintenance model. The proposed model solves the CMS along with security-constrained unit commitment (SCUC) over 1-year planning horizon and the final SCUC solution prescribes an economic and secure operation for composite power systems. The method is demonstrated on the IEEE 30-bus, IEEE reliability test system (single and three areas) and practical Indian utility 19-unit power systems, and the results are presented.

Economic power dispatch with cubic cost models using teaching learning algorithm

Economic dispatch (ED) solution accuracy can be improved with cubic cost models and optimisation algorithms. This article proposes a new methodology for solving ED problem with cubic cost models using teaching learning-based optimisation (TLBO) algorithm. The key aspects of ED scenario such as valve point effects, environmental factors, transmission losses, spinning reserve, ramp rate, prohibited operating zones and fuel limitations are considered in this study. The proposed methodology is applied to test systems involving cubic cost equations in 3, 5, 6, 13, 26 and a large-scale system containing 156 units, in order to evaluate its efficiency and feasibility. Convergence characteristics of the TLBO has been assessed and investigated through comparison with results reported in the literature. Many trials with different initial values have been carried out for all the test systems in order to justify the robustness of the proposed methodology. Considering the quality of the solution and convergence speed obtained, this method seems to be a promising alternative approach for solving the ED problems with cubic functions.

Use of a multi-objective teaching-learning algorithm for reduction of power losses in a power test system

This paper presents a multi-objective teaching learning algorithm based on decomposition for solving the optimal reactive power dispatch problem (ORPD). The effectiveness and performance of the proposed algorithm are compared with respect to a multi-objective evolutionary algorithm based on decomposition (MOEA/D) and the NSGA-II. A benchmark power system model is used to test the algorithms’ performance. The results of the power losses reduction as well as the performance metrics indicate that the proposed algorithm is a reliable choice for solving the problem.

Modified teaching learning algorithm and double differential evolution algorithm for optimal reactive power dispatch problem: A comparative study

One of the major optimization problems in power systems is the optimal reactive power dispatch (ORPD) problem. ORPD problem is a multi-variable problem with non-linear characteristics and has established significant emphasize on secure and economical operation of power systems. Therefore, finding an effective algorithm to solve this problem is absorbing a growing attention among the academics. In this paper, a reliable and effective algorithm is introduced by combining modified teaching learning algorithm (MTLA) and double differential evolution (DDE) algorithm in order to handle the ORPD problem. Moreover, the proposed algorithm is applied to ORPD problem on IEEE 14-bus, IEEE 30-bus and IEEE 118-bus power systems for performance assessment and validation purposes. The obtained results demonstrate the efficiency of the proposed hybrid algorithm and show faster convergence and better solutions in comparison with previously reported algorithms.

A novel hybrid algorithm of imperialist competitive algorithm and teaching learning algorithm for optimal power flow problem with non-smooth cost functions

One of the major tools for power system operators is optimal power flow (OPF) which is an important tool in both planning and operating stages, designed to optimize a certain objective over power network variables under certain constraints. Without doubt one of the simple but powerful optimization algorithms in the field of evolutionary optimization is imperialist competitive algorithm (ICA); outperforming many of the already existing stochastic and direct search global optimization techniques. The original ICA method often converges to local optima. In order to avoid this shortcoming, we propose a new method that profits from teaching learning algorithm (TLA) to improve local search near the global best and a series of modifications is purposed to the assimilation policy rule of ICA in order to further enhance algorithm’s rate of convergence for achieving a better solution quality. This paper investigates the possibility of using recently emerged evolutionary-based approach as a solution for the OPF problem which is based on hybrid modified ICA (MICA) and TLA (MICA–TLA) for optimal settings of OPF control variables. The performance of this approach is studied and evaluated on the standard IEEE 30-bus and IEEE 57-bus test systems with different objective functions and is compared to methods reported in the literature. The hybrid MICA–TLA provides better results compared to the original ICA, TLA, MICA, and other methods reported in the literature as demonstrated by simulation results.

Reactive Power Handling by a Multi-Objective Teaching Learning Optimizer Based on Decomposition

The teaching learning-based optimization (TLBO) is a population-based optimization algorithm suitable for solving complex problems. TLBO imitates the interaction between a teacher and her/his students. The global solution search process of this approach consists of two phases: the teacher- and the learner-phase. This paper proposes a multi-objective teaching learning algorithm based on decomposition (MOTLA/D) for solving a reactive power handling problem. The proposed method is validated on three test systems, and it is compared with respect to a state-of-the-art multi-objective evolutionary algorithm based on decomposition (MOEA/D).

A new modified teaching-learning algorithm for reserve constrained dynamic economic dispatch

This paper presents a new optimization algorithm, named modified teaching-learning algorithm, to solve a more practical formulation of the reserve constrained dynamic economic dispatch of thermal units considering the network losses and operating limitations of the generating units (i.e., the valve loading effect and ramp rate limits). Unlike the previous approaches, three types of the system spinning reserve requirements are explicitly modeled in the problem and a new constraint-handling is proposed to satisfy them. The proposed teaching-learning optimization algorithm is a new population-based optimization method features between the teacher and learners (students). Therefore, this algorithm searches for the global optimal solution through two main phases: 1) the “teacher phase” and 2) the “learner phase”. Nevertheless, these two phases are not adequate for learning interaction between the teacher and the learners in the entire search space. Thus, in this paper a new phase named “modified phase” based on a self-adaptive learning mechanism is added to the algorithm to improve the process of knowledge learning among the learners and accordingly generate promising candidate solutions. The proposed framework is applied to 5-, 10-, 30-, 40-, and 140-unit test systems in order to evaluate its efficiency and feasibility.

Probabilistic multiobjective wind-thermal economic emission dispatch based on point estimated method

In this paper, wind power generators are being incorporated in the multiobjective economic emission dispatch problem which minimizes wind-thermal electrical energy cost and emissions produced by fossil-fueled power plants, simultaneously. Large integration of wind energy sources necessitates an efficient model to cope with uncertainty arising from random wind variation. Hence, a multiobjective stochastic search algorithm based on 2 m point estimated method is implemented to analyze the probabilistic wind-thermal economic emission dispatch problem considering both overestimation and underestimation of available wind power. 2 m point estimated method handles the system uncertainties and renders the probability density function of desired variables efficiently. Moreover, a new population-based optimization algorithm called modified teaching-learning algorithm is proposed to determine the set of non-dominated optimal solutions. During the simulation, the set of non-dominated solutions are kept in an external memory (repository). Also, a fuzzy-based clustering technique is implemented to control the size of the repository. In order to select the best compromise solution from the repository, a niching mechanism is utilized such that the population will move toward a smaller search space in the Pareto-optimal front. In order to show the efficiency and feasibility of the proposed framework, three different test systems are represented as case studies.