Blended acquisition along with efficient detector and source geometries allows for a cost-effective operation. The outcome of subsequent deblending and data reconstruction is of primary importance in determining the technical success of this manner of data acquisition. Despite its advantages over conventional seismic surveys, finding optimum survey parameters is a difficult task. Although incorporating irregularity into spatial sampling and blending schemes leads to effective deblending and data reconstruction, it inherently provides a significantly large problem space. We have developed a survey-design workflow to provide acquisition parameters that account for the source blending as well as the spatial sampling of detectors and sources in an automated manner. Our method involves an iterative scheme to derive the survey parameters that lead to optimum deblending and data reconstruction quality. The approach deals jointly with deblending and data-reconstruction via a sparse inversion in the frequency-wavenumber domain coupled with constraints based on causality and coherency. The residue from this process is subsequently used to update the survey parameters by integrating a genetic algorithm and a convolutional neural network (CNN). Bioinspired operators enable the simultaneous update of the blending and sampling operators. To relate the choice of survey parameters to the performance of deblending and data reconstruction, we have used a CNN. The applied network architecture successfully rejects suboptimal solutions among newly generated ones from genetic operators. Consequently, only optimal ones are fed into the subsequent step, making our approach computationally affordable. We apply our workflow to design a seismic survey that incorporates the dispersed source array concept. A comparison among different survey-design strategies highlights the ability of the method to effectively derive optimum solutions. The resultant acquisition scenario derived from our approach yields a notable enhancement of deblending and data reconstruction quality attributed solely to the choice of survey parameters.