Walid Ellili - Academia.edu (original) (raw)
Papers by Walid Ellili
Through the present paper, a new approach useful for solving the obstacle avoidance and trajector... more Through the present paper, a new approach useful for solving the obstacle avoidance and trajectory optimization problems during robot navigation for certain tasks to be performed at minimum costs. In its real sense, the obstacle avoidance approach is based on the application of two fuzzy controllers, the first of which is designed to join the object, while the second is conceived to serve as an obstacle avoiding device. The trajectory optimization approach is based on the gradient method. Prior to implementing the solution on the real robot, the simulation has been integrated in an immersive virtual environment, for a more effective movement analysis and safer testing purposes. The study findings prove to reveal well that the proposed approach turns out to exhibit a good average speed and a satisfactory target-reaching success rate, while the optimization oriented gradient method has turned out to be rather efficient in respect of the genetic algorithm approach.
The present work is proposed to establish a comparative study between the exact method, namely, t... more The present work is proposed to establish a comparative study between the exact method, namely, the Algorithm Branch and Bound, in respect of the artificial intelligence approaches (the genetic algorithms as well as the neural ones). The purpose is to highlight the artificial intelligence noticeable performance in solving the multi-product and multi-period model through the abovementioned method, with the aim of providing a solution for the reverse logistics' site-localization problem regarding end-of-life products. For the sake of reaching, a solution within a reasonable time, however, the genetic algorithm and neural network have displayed a remarkable ability to effectively solve the problem, as considered in relation to the assessment and sorting or separation procedure (branch and bound algorithm), constructed within a CPLEX shopping solver. In addition, a comparative study between the three methods is going to be established.
Traveling Salesman Problem consists in finding the shortest distance commercial representatives s... more Traveling Salesman Problem consists in finding the shortest distance commercial representatives should undertake on visiting the 24 cities in Tunisia prior to resuming the initial departure point. Such a deceptive issue appears to stand as a remarkable challenge in computational mathematics. The purpose of this paper lies in implementing genetic Mat lab's algorithms toolbox gads in a bid to cope with such a problem.
Through the present paper, a new approach useful for solving the obstacle avoidance and trajector... more Through the present paper, a new approach useful for solving the obstacle avoidance and trajectory optimization problems during robot navigation for certain tasks to be performed at minimum costs. In its real sense, the obstacle avoidance approach is based on the application of two fuzzy controllers, the first of which is designed to join the object, while the second is conceived to serve as an obstacle avoiding device. The trajectory optimization approach is based on the gradient method. Prior to implementing the solution on the real robot, the simulation has been integrated in an immersive virtual environment, for a more effective movement analysis and safer testing purposes. The study findings prove to reveal well that the proposed approach turns out to exhibit a good average speed and a satisfactory target-reaching success rate, while the optimization oriented gradient method has turned out to be rather efficient in respect of the genetic algorithm approach.
The introduction of reverse logistics has been remarkably enhanced by the relevant legislative ac... more The introduction of reverse logistics has been remarkably enhanced by the relevant legislative acts passed in several industrial countries, with the aim of preserving the environment. Given the diverse activities and definitions attributed to this particular type of logistics, the relevant design and control turn out to be too sophisticated. It is in this particular context that the present paper can be set, with the major objective of implementing a mixed integer nonlinear programming model for the design of an integrated distribution network, which appears to be dynamic once the integrated nature of reverse logistics network optimization is being considered. As demonstrated, the genetic algorithms' effectiveness in achieving the most optimum solutions within a reasonable time framework has also been highlighted. In a last stage, these methods' limitations have been underlined, along with some suggested research perspectives. An analysis of some mathematical reverse logistics models has also been undertaken with respect to five related application areas, namely: location, life cycle assessment, production planning, inventory management, along with the establishment of the most appropriate product-collection routes. In addition, the reverse logistics' case studies, as treated in the pertinent literature, have also been thoroughly analyzed and grouped according to industries.
Traveling Salesman Problem consists in finding the shortest distance commercial representatives s... more Traveling Salesman Problem consists in finding the shortest distance commercial representatives should undertake on visiting the 24 cities in Tunisia prior to resuming the initial departure point. Such a deceptive issue appears to stand as a remarkable challenge in computational mathematics. The purpose of this paper lies in implementing genetic Mat lab's algorithms toolbox gads in a bid to cope with such a problem.
14th International Conference on Sciences and Techniques of Automatic Control & Computer Engineering - STA'2013, 2013
ABSTRACT To preserve the environment, a number of industrial countries is established legislative... more ABSTRACT To preserve the environment, a number of industrial countries is established legislative laws that encourage the development of the diversity of activities and the definitions given in this type of logistics, its design and its management are very complicated. We have demonstrated the efficacy of neural network tried using the programming language level MATLAB; generate a database containing all the necessary data to ensure learning neural networks, reverse logistics. Given works to obtain the best solutions.
Through the present paper, a new approach useful for solving the obstacle avoidance and trajector... more Through the present paper, a new approach useful for solving the obstacle avoidance and trajectory optimization problems during robot navigation for certain tasks to be performed at minimum costs. In its real sense, the obstacle avoidance approach is based on the application of two fuzzy controllers, the first of which is designed to join the object, while the second is conceived to serve as an obstacle avoiding device. The trajectory optimization approach is based on the gradient method. Prior to implementing the solution on the real robot, the simulation has been integrated in an immersive virtual environment, for a more effective movement analysis and safer testing purposes. The study findings prove to reveal well that the proposed approach turns out to exhibit a good average speed and a satisfactory target-reaching success rate, while the optimization oriented gradient method has turned out to be rather efficient in respect of the genetic algorithm approach.
The present work is proposed to establish a comparative study between the exact method, namely, t... more The present work is proposed to establish a comparative study between the exact method, namely, the Algorithm Branch and Bound, in respect of the artificial intelligence approaches (the genetic algorithms as well as the neural ones). The purpose is to highlight the artificial intelligence noticeable performance in solving the multi-product and multi-period model through the abovementioned method, with the aim of providing a solution for the reverse logistics' site-localization problem regarding end-of-life products. For the sake of reaching, a solution within a reasonable time, however, the genetic algorithm and neural network have displayed a remarkable ability to effectively solve the problem, as considered in relation to the assessment and sorting or separation procedure (branch and bound algorithm), constructed within a CPLEX shopping solver. In addition, a comparative study between the three methods is going to be established.
Traveling Salesman Problem consists in finding the shortest distance commercial representatives s... more Traveling Salesman Problem consists in finding the shortest distance commercial representatives should undertake on visiting the 24 cities in Tunisia prior to resuming the initial departure point. Such a deceptive issue appears to stand as a remarkable challenge in computational mathematics. The purpose of this paper lies in implementing genetic Mat lab's algorithms toolbox gads in a bid to cope with such a problem.
Through the present paper, a new approach useful for solving the obstacle avoidance and trajector... more Through the present paper, a new approach useful for solving the obstacle avoidance and trajectory optimization problems during robot navigation for certain tasks to be performed at minimum costs. In its real sense, the obstacle avoidance approach is based on the application of two fuzzy controllers, the first of which is designed to join the object, while the second is conceived to serve as an obstacle avoiding device. The trajectory optimization approach is based on the gradient method. Prior to implementing the solution on the real robot, the simulation has been integrated in an immersive virtual environment, for a more effective movement analysis and safer testing purposes. The study findings prove to reveal well that the proposed approach turns out to exhibit a good average speed and a satisfactory target-reaching success rate, while the optimization oriented gradient method has turned out to be rather efficient in respect of the genetic algorithm approach.
The introduction of reverse logistics has been remarkably enhanced by the relevant legislative ac... more The introduction of reverse logistics has been remarkably enhanced by the relevant legislative acts passed in several industrial countries, with the aim of preserving the environment. Given the diverse activities and definitions attributed to this particular type of logistics, the relevant design and control turn out to be too sophisticated. It is in this particular context that the present paper can be set, with the major objective of implementing a mixed integer nonlinear programming model for the design of an integrated distribution network, which appears to be dynamic once the integrated nature of reverse logistics network optimization is being considered. As demonstrated, the genetic algorithms' effectiveness in achieving the most optimum solutions within a reasonable time framework has also been highlighted. In a last stage, these methods' limitations have been underlined, along with some suggested research perspectives. An analysis of some mathematical reverse logistics models has also been undertaken with respect to five related application areas, namely: location, life cycle assessment, production planning, inventory management, along with the establishment of the most appropriate product-collection routes. In addition, the reverse logistics' case studies, as treated in the pertinent literature, have also been thoroughly analyzed and grouped according to industries.
Traveling Salesman Problem consists in finding the shortest distance commercial representatives s... more Traveling Salesman Problem consists in finding the shortest distance commercial representatives should undertake on visiting the 24 cities in Tunisia prior to resuming the initial departure point. Such a deceptive issue appears to stand as a remarkable challenge in computational mathematics. The purpose of this paper lies in implementing genetic Mat lab's algorithms toolbox gads in a bid to cope with such a problem.
14th International Conference on Sciences and Techniques of Automatic Control & Computer Engineering - STA'2013, 2013
ABSTRACT To preserve the environment, a number of industrial countries is established legislative... more ABSTRACT To preserve the environment, a number of industrial countries is established legislative laws that encourage the development of the diversity of activities and the definitions given in this type of logistics, its design and its management are very complicated. We have demonstrated the efficacy of neural network tried using the programming language level MATLAB; generate a database containing all the necessary data to ensure learning neural networks, reverse logistics. Given works to obtain the best solutions.