Networked infrastructures such as gas and water pipelines, roads, railroads or power grids provide essential utilities and services to society. Common characteristics of such infrastructures include high capital costs, generally long lifetimes and irreversibility once the construction of such networks have finished. In the design of these networks, the planners face a multitude of challenges ranging from traditional factors such as technical complexities, space-constrained areas to emerging factors such as complex multi-actor contexts and climate change. A chronic challenge is a multi-actor context in which supply capacities from the supplier side, demand capacities from the consumer side and information about the actual commitment of the network participants, who are about to be connected to the network, can remain uncertain for a long time. The uncertainty is especially high when the network involves multiple suppliers and multiple consumers. While deterministic network design ignores these uncertainties, non-deterministic network design takes them into consideration. The goal of this research is to develop an approach of designing a multiple source (supplier) - multiple sink (consumer) network layout that minimizes the initial investment costs while remaining flexible in its response to future changes of network participants. To this end, firstly an agent-based deterministic modeling method of Ant Colony Optimization was developed, which proved to be feasible in finding cost-minimized network layouts of multiple sources and multiple sinks. Next, the method was extended to a non-deterministic method by embedding flexibility options in order to deal with the uncertainty on network participants. The modeling methods of Ant Colony Optimization were found to be intuitive, extensible and customizable. Based on the modeling outputs, the final design approach can be a supportive decision-making tool in the network planning stage. Future work needs to incorporate more practical criteria required by decision makers into the modeling to bridge the gap between scientific knowledge and decision making process in reality.