DEVELOPMENT OF SOIL TYPE REFERENCE OF METROPOLITAN MANILA USING GENETIC ALGORITHM

Abstract

A soil type reference of Metropolitan Manila was developed using an optimization algorithm called Genetic Algorithm (GA). In this study, a program was created to perform a Genetic Algorithm on each meter soil layer ranging from 40 meters below mean sea level up to 100 meters above means sea level for specified points with 2km grid intervals. The program utilized the Unified Soil Classification System (USCS) and a borehole database of the study area and surrounding provinces to predict the soil types. Two variables were chosen to assess the fitness of the prediction: the likeness of the soil type with surrounding boreholes, and the distance of the boreholes from a specified grid point. A total of 216 points were used in the study each containing 40 to 130 soil layers. The results were then compiled and plotted to create a total of 38 soil profiles of the study area for the longitudinal and transverse directions. The profiles followed the expected soil types based on the geologic zones of the study area: sands and clays for the coastal lowlands, rocks and sands for the central plateau, and clays and silts for Marikina Valley. Validation of the profiles suggests the Genetic Algorithm as a suitable tool for the development of soil type reference of the study area.