The antsystem (AS) algorithm of Dorigo is a new computational paradigm, which is a stochastic combinatorial algorithm. It solves optimization problems by means of "ants", that is, agents with a very simple ...
详细信息
ISBN:
(纸本)0780370872
The antsystem (AS) algorithm of Dorigo is a new computational paradigm, which is a stochastic combinatorial algorithm. It solves optimization problems by means of "ants", that is, agents with a very simple basic capability, which mimic the behavior of real ants. The AS proposed by Dorigo has appealing features, but in its standard form it has some limitations. Applied to the Traveling Salesman Problem, the AS approach encounters difficulties when applied to random graphs. To remedy this, we design a new type of agent by using intensification and diversification strategies, based on the proposals of tabu search, in order to reach better solutions.
Classic algorithms that are based on the antsystem theory have been designed to face problems with discrete solutions. When dealing with non discrete problems, in order to apply ant Colony Optimization (AGO) algorith...
详细信息
Classic algorithms that are based on the antsystem theory have been designed to face problems with discrete solutions. When dealing with non discrete problems, in order to apply ant Colony Optimization (AGO) algorithms they should be transformed to discrete problems. The arbitrary limitation of the number of possible solutions for each space is the result of the transformation of the solutions of the problems from the continuous to the discrete workspace. Hence, the choice of the width of the solutions spaces essentially defines the possible best solutions of the problem. In order to deal with this disadvantage, we present a new algorithm: the Continuous ant Colony (C-ant). This algorithm encourages local searching around the best solution found in each iteration. The proposed (C-ant) is applied to a simple ELD problem composed of 4 generators. Comparison to conventional Particle Swarm Optimization (PSO) algorithms is presented.
作者：
Rui YangEAIC Dept
China Waterborne Transportation Institute Beijing China
This paper aimed to optimize the technique parameters of tracking buoy by a hydrodynamic method through an experiment on the sea. A survey was created based on the study of technical characters of oil spill tracking b...
详细信息
This paper aimed to optimize the technique parameters of tracking buoy by a hydrodynamic method through an experiment on the sea. A survey was created based on the study of technical characters of oil spill tracking buoy to achieve an allweather whole procedure monitoring propose for oil spill by using of satellite positioning communication mode, which can provide an effective technical method for the rapid response of oil spill emergency.
Intrinsically, the attainment of optimal solution via the ant Colony system (ACS) algorithm essentially depends on the attractiveness of the quantity of pheromone on a given path. This leads to neglecting the velocity...
详细信息
Intrinsically, the attainment of optimal solution via the ant Colony system (ACS) algorithm essentially depends on the attractiveness of the quantity of pheromone on a given path. This leads to neglecting the velocity of the ant which constitutes an important nature-based heuristic information. The aim of this paper is to improve an existing ACS algorithm by integrating ant velocity, an insight gained from the Intelligent Water Drops (IWD) algorithm. A bi-objective model was formulated and adapted into the proposed ACS algorithm to optimize route length and social cost associated with various activities along the route. The solution technique was based on the min-max approach. A 14-node road network data, measuring distances and social costs was used in validating the algorithm. Both the benchmark algorithm and our proposed ant velocity-based ACS algorithm yielded the same bi-optimal solution (12 km, GHS 7) of distance and social cost along the path 1 -> 4 -> 7 -> 11 -> 12 -> 14. The proposed ACS algorithm converges at the 127th iteration, corresponding to approximately 3 s execution time. Obviously, the proposed ACS algorithm outperforms the benchmark algorithm which converges at the 207th iteration, with approximately 5 s execution time. Therefore, the proposed ACS algorithm has outperformed the benchmark ACS algorithm in respect of time (or the number of iterations needed for convergence) by approximately 39 %. Evidently, with a velocity of 0.2445 ms-2, the optimal time taken by the best ant to complete the tour is approximately 27 s.
暂无评论