A fingerprinting algorithm based on artificial intelligence genetic algorithm

Abstract

Oil spill identification is an important means of investigation of oil spill through analysis and comparison of suspicious oil spill source. This paper proposed a GC/MS fingerprint fitting method based on genetic algorithm (GA) to identify the source of oil spill. In this method, dominant characteristic peaks were selected from the 3d fingerprint of the standard and stored in the prevailing characteristic fingerprint database (dominant database). Recessive characteristic peaks were screened out from the periphery of the dominant characteristic peaks of the sample and stored in the recessive characteristic fingerprint database (recessive database). Then a set of characteristic peaks was selected from the dominant library by GA to correct the characteristic peaks in the samples and automatically identify the similarity between the standard and the samples. Simulation calculation compared with experimental analysis results showed that the method to build a detailed identification of oil spill fingerprint was accurate and reliable. The method to identify the fingerprint peak can be clearly and accurately reflect the properties of the sample, which can be used for rapid automatic calculation of chromatographic fingerprint similarity, so as to strengthen sea spill accident discriminant scientific theory in identification, traceability and the application of legal liability.