A Novel Artificial Bee Colony Algorithm with an Overall-Degradation Strategy and Its Performance on the Benchmark Functions of CEC 2014 Special Session
Issue:
Volume 2, Issue 5, October 2014
Pages:
71-80
Received:
23 January 2014
Accepted:
4 September 2014
Published:
30 September 2014
Abstract: The artificial bee colony (ABC) algorithm has been a well-known swarm intelligence algorithm, which assimilates the cooperating behavior of bees when seeking for nectar sources. Aiming to improve the conventional ABC algorithm, we focus on the re-initialization phase. In this paper, an overall-degradation-oriented artificial bee colony (OD-ABC) algorithm is proposed, pursuing to fight against premature convergence. This is achieved through re-initializing majority of the employed bees at one time, rather than generating at most one scout bee in each iteration. In this work, our OD-ABC algorithm is compared against the conventional ABC algorithms using 24 benchmark functions that origin from the CEC 2014’s competition on single objective real-parameter numerical optimization. The numerical results show that the OD-ABC algorithm is effective and thus can be employed to fight against premature convergence.
Abstract: The artificial bee colony (ABC) algorithm has been a well-known swarm intelligence algorithm, which assimilates the cooperating behavior of bees when seeking for nectar sources. Aiming to improve the conventional ABC algorithm, we focus on the re-initialization phase. In this paper, an overall-degradation-oriented artificial bee colony (OD-ABC) alg...
Show More
A Discrete-Time Quasi-Theoretical Solution of the Modified Riccati Matrix Algebraic Equation
Issue:
Volume 2, Issue 5, October 2014
Pages:
87-92
Received:
31 October 2014
Accepted:
7 November 2014
Published:
20 November 2014
Abstract: In this paper, based on MacLaurin’s series and the Riccati equation, an algebraic quadratic equation will be developed and hence, its two roots, which represent the minimizing and maximizing optimal control matrices, would be deducted easier. Otherwise, a step-by-step algorithm to compute the control matrix for every step of time according to the preceding responses and a new signal pick will be explained. The proposed method presents a new discrete-time solution for the problem of optimal control in the linear or nonlinear cases of systems subjected to arbitrary signals. As an example, a system (structure) of three degrees of freedom, subjected to a strong earthquake is analyzed. The displacements versus time and the stiffness forces versus displacements of the system, for the two uncontrolled and controlled cases are graphically shown. Therefore, the curves of variations of the elements of the optimal control matrix versus discrete-time are also presented and clearly show the effect of the nonlinearity, of the system, which is the cause of the great responses in the uncontrolled case, and that it is optimally treated by the proposed solution. The results obtained clarify a great reduction of the controlled system results, in comparison with the uncontrolled system ones. The percentage of the differences between the controlled and uncontrolled results (displacements or stiffness forces) could even surpass 90 %, which demonstrates that the adopted solution is good even than that of the original ones of the differential or the algebraic Riccati equation.
Abstract: In this paper, based on MacLaurin’s series and the Riccati equation, an algebraic quadratic equation will be developed and hence, its two roots, which represent the minimizing and maximizing optimal control matrices, would be deducted easier. Otherwise, a step-by-step algorithm to compute the control matrix for every step of time according to the p...
Show More
A Modified Kneed Biped Real Robot Based on Parametric Excitation Principle
Yoshihisa Banno,
Kouichi Taji,
Yuji Harata,
Kyohei Seta
Issue:
Volume 2, Issue 5, October 2014
Pages:
93-99
Received:
5 November 2014
Accepted:
15 November 2014
Published:
20 November 2014
Abstract: Parametric excitation walking is one of the bipedal gait generation methods on level ground. This method was first applied to a biped robot with telescopic legs and later to a kneed biped robot. An experimental robot with telescopic legs was also developed and it was verified that the robot could walk more than eight steps by the parametric excitation walking. Recently, we have developed an experimental kneed biped robot and have shown the robot can walk more than fifteen steps stably in inverse bending fashion. But the robot has a deficiency in that the robot does not have a ground sensor and the robot is controlled only in open-loop fashion. In this paper, we modify and improve the robot by using a ground sensor and shock absorbing material to enable to control in closed-loop fashion and hence, to improve the gait performance. The experiments are performed and the walking performance of the robot is investigated. The experimental results are compared with the numerical results, and the validity of the numerical simulation is verified.
Abstract: Parametric excitation walking is one of the bipedal gait generation methods on level ground. This method was first applied to a biped robot with telescopic legs and later to a kneed biped robot. An experimental robot with telescopic legs was also developed and it was verified that the robot could walk more than eight steps by the parametric excitat...
Show More
