Lottery drawing

Abstract

We consider the problem of an unbiased choice of k items at random form a total of n candidates, where k<=n,This problem has many applicatioins, for example,in quality control or lottery drawing. We present three algorithms to tackle the problem. All three algorithms have their own merit and are based on using random numbers. Due to the nondeterministic characteristics, the execution time for each algorithm is primarily determined by the number of random numbers used it its process.We formally analyze each algorithm by computing its expected number of random numbers required and the variance associated with it. Computer simulations are used to gain further insights of these algorithms. FInally, we present a general guide-line to direct users in selecting an appropriate algorithm based on the values of k and n.