This study explores to what extent Agent Based Systems (ABS) can handle multi-criteria optimization problems. The implementation of ABS in the field of optimization has limitations to address multiple criteria in a continuous generation process due to ABS usually merge the perceived information in a single response. To address this limitation, we increase the responsiveness of the systems through a multiple production approach. This approach breaks down the problem into two parts: the configuration through the interactions of the agents, and the overall performance through their local decisions. The method is tested in a case study of the network circulations of a park, optimizing the slope, views and sun. Performance and differentiation capabilities are evaluated in populations generated in two different scenarios. Data analysis methods verify the effectiveness of the algorithm and quantify the influence of each parameter on the final results.