CONSTRUCTION OF SAFETY EVALUATION MODEL FOR TAIWAN CHILDREN'S PLAYGROUND EQUIPMENT USING THE HIERARCHICAL BAYESIAN APPROACH AND GENETIC ALGORITHMS—CASE STUDIES OF SWING AND SLIDE

Abstract

ABSTRACT “Playing Games,” one of the key parts of a child's life, is essential to their growth and learning. In recent years, negligence and misuse of children's playground equipment, many with defects in design, have caused many cases of injury during play. On the average, one child dies each day due to accident related injuries. According to analysis of injury and safety management related information, the occurrence of accidents involving play areas has a direct relationship to the proper and improper use of such equipment. To fully understand the actual relationship between the design of playground equipment and accident related injuries during use, we construct a model of the relationship of the degree of injuries sustained by children (i.e. flesh wounds, serious wounds) in this study. For this model, we have chosen the design of the most common playground facilities: the swing and slide, as our research target, and have selected elementary students from each city and county in Taiwan between five to twelve years old as research objects. During the construct of this model, we shall primarily use the Hierarchical Bayesian Approach as the main reference when analyzing the data. Then, by using a simplified data analyzing technique—the Hierarchical Bayesian Model, we shall calculate the parameters using the research model. After finding the main factors of product designs that affects children injuries, the model will be provided as a reference to the toy industry to the future design of playground equipment. In addition, in order to resolve the efficiency issues regarding the reference estimates of the Hierarchical Bayesian model, we have integrated the genetic algorithm (GA) technology to expedite the speed of convergence amid the Monte Carlo Markov Chain (MCMC). Based on actual application results, we have discovered that the proposed method contains obvious supportive functions.