new weighting strategies for correlated poisson sampling

In poisson sampling, the units are sampled independently with given inclusion probabilities and the sample size n(s) is a random variable so poisson sampling gives a large variety of the sample size. also, if there exists a trend in the y which poisson sampling cannot take into account. but correlated poisson sampling is an alternative to poissonsampling that uses weights to make correlations between the inclusion indicators. in that manner, it is possible to decrease the variety of the sample size and it is even likely to have a constant sample size to make the samples more evenly spread over the population. in this paper, new strategies for choosing the weights are investigated by simulation, and willbe shown that the new strategies increase the efficiency of the estimated parameter.