| A new type of simulated evolutionary algorithm, ant colony algorithm, is researched in this paper, which is used to solve some NP-hard combinatorial optimization problem. The Ant Colony Optimization (ACO) is a relatively new meta-heuristic algorithm, The ACO solves problem through mimicking ants' foraging behavior. It is the example that relatively succeeds in the bionics algorithm at present.Though bionics algorithms such as Genetic Algorithms, Neural Network, etc. solve and make the problem of optimizing up at present, and has already obtained very good development, but been regarded as the new member of the bionics family—ACO, optimize with the overall situation, positive feedback, distributed computing, intelligence search, apt characteristic of combining with other algorithms etc. has been paid close attention to the fact by the people more and more.This algorithm was proposed in 1991 by M. Dorigo, etc. first, up until now, through the research of more than ten years, it has already obtained the development at full speed. ACO as one principle simple new bionic, have already been applied to a lot of kinds of fields, and has applied to the question of businessman in the travel industry and obtained great success , thereafter, applied to the question of QAP , Job-shop , GCP and VRP ,etc. which also have made very good result. With the deepening studying, a lot of scholars have proposed some improved optimization algorithms in succession, for instance Rank-Based Version of Ant System, (ASrank) Ant Colony System, (ACS), Max-Min Ant System (MMAS) , etc.This paper makes one summary about the development of algorithm in the past at first. Then propose a new kind of improvement algorithm-ACOrw secondly. And test the feasibility of examining the algorithm of two angles from the theory and computer emulation separately. At last, by using extension's excellent degree appraised carry on comprehensive appraisal of the algorithm-ACO, an improved ACO and ACOrw.
|
PDF Full Text Request