Study Of Multi-scale Niche Ant Colony Algorithm For Parameters Inversion Of The Modified BISQ Model

Seismic wave exploration technique is one of the most widely used methodsin energy resources exploration and development. Main energy resources areburied in a complex geological structure and physical properties of the earth’slithosphere. Actually, we often simplified the complex geological structure whichcontain resources as geometric model for simulation during exploration. Butafter simplifying the process of model, there has an error with actual situationwhich makes the results of numerical simulation inaccurate. A high precisionresearch about the geological exploration field is needed, the hypothesis which ismore accordance with the actual geological structure is urgently to be solved.The modified BISQ model is based on the original BISQ model. The maindifference between the modified BISQ model and the original BISQ model is thatthe former has a more simple calculation formula and can relatively reducenumerical computation. As a new global optimization algorithm in the field ofoptimization algorithm, Ant Colony Algorithm has many advantages, forexample, using distributed parallel computing system, easily to be combinedwith other algorithms, has wide applicability and good capability of globaloptimization. But it also needs a longer time to research and easily to fall intolocal optimal solutions at the same time. In view of the shortcomings oftraditional ant colony algorithm, researchers put forward many improvements.Multi-scale niche ant colony algorithm is based on multi-scale ideas aftercombined with niche ant colony algorithm. First, we applied the algorithm to themultimodal function optimization problems and compared the results with theniche ant colony algorithm to test and verify the effectiveness. Second, use themulti-scale niche ant colony algorithm in single and double parameter’sinversion based on the modified BISQ model. It will get a good result when theinitial parameter range is expanded compared to the niche ant colony algorithm’sresult. But the multi-scale niche ant colony algorithm should have a improvementin the inversion time. Finally, we consider the case of adding disturbance whenusing the multi-scale niche ant colony algorithm to inversion (add5%noise and10%noise). The results show the precision is declining when the disturbance isincreasing, but the whole precision is desirable. The relative error of penetrationis under0.09and solid density is under0.01. The high precision results verifythe feasibility and stability of the algorithm.The multi-scale niche ant colony algorithm in this dissertation is the firsttime we combine the multi-scale ideas with niche ant colony algorithm. The precision of the parameters inversion is very high. but the algorithm is notperfect, we hope it will be modified with the development of research.