The Study Of Competitive Decision-making Algorithm For Online Currency Trading

On-line problem and its competitive algorithm theory are hot research directions in recent years at home and abroad. On-line problem is the study of decision-making under the incomplete information conditions, Since it is impossible to know and predict the future information exactly, people often can not make the optimal decision on the on-line problem but only can make a satisfied decision as much as possible. The competitive algorithm is one such decision strategy, the decision results of the competitive decision algorithm under various conditions are within a certain range corresponding to the offline optimal decision. The statement of the online competition algorithms and the concept of competitive ratio make up for the insufficients of the traditional optimization theory in dealing with the online problems. Many research achievements have been acquired in the computer science and it obtains the widespread applications.In recent years, along with the globalization of trade and the acceleration of economic integration process, and because of the lack of future information and uncertainty in financial and management problems, now they are also increasingly received numerous online problems and competitive algorithm researcher's extensive attention. Online trading algorithms not only have the extremely important theoretical significance in the economic field, but also have the wide range of applications in economic management and the daily life, such as search employees and work, stock investment,inventory and insurance issues and so on. In reality, there are a variety of different models of the typical online trading decision-making have been solved by deterministic algorithms and random algorithms. This article describes the basic model of online treading problem and related deterministic and stochastic algorithms, and propose an random decision algorithm RAKD(Random Algorithm with Known Duration) based on the time-series search model that introduced profit function, analysis and prove the worst case competitive ratio, finally compare it with deterministic algorithm on algorithm performance through the experiments, the analysis and the experimental results show that the algorithm can reduce the competitive ratio under certain circumstances, and application scope are more widely. In addition, the article gives the improved random algorithm to the non worst case.Finally, the paper summarizes the research work, and points out that what need to be further studied in the area by online trading decision algorithms.