The contributions of this paper is adopting the Support Vector Regression Guided Genetic Algorithm (SVRGGA) presented as an improved version of traditional Genetic Algorithm (GA), which is inspired from the individual self-adaptive capability observed in nature with additional guidance information from a statistical response of cost distributions conditioned on the locations of facilities. In the proposed method, we used SVRGGA to solve the location problem and internal GA to challenge allocation issue. Since location and allocation are not two separate problems, we must solve the two problems simultaneously. Therefore, we combined the two genetic algorithms in one algorithm in such a manner that output of one is used as input to another.
Indeed, each individual solution is empowered in a GA population with the capability of self- adjustment to boost the fitness for the objective function. This is inspired by the observation of an individual's self-adaptivity to meet the requirement of the environment for many species in the natural world. In the SSCMFWP context, SVRGGA was proposed to allow the individuals in the conventional GA to adjust its fitness with the guiding information from a probability distribution aggregated from SVR surface[2, 4] and the Monte-Carlo simulation [3].