So, I too have been trying to accomplish this.
From what I've gathered, particles and fragments do not act as attracting bodies. A shame, but makes sense from a simulation performance stand point. However, I have been using a method that offers acceptable results.
To "birth" a system, you can try the following:
-Add a star or any object with mass to be the center. (Optional)
-Use the Add Object -> Rings, select "Asteroid Belt" (personal preference), and edit the parameters.
+ Choose how many particles you would like to generate of a specific material type.
+ Choose the cumulative mass of all the particles. Keep in mind the mass of each particle after generation.
+ Select which material you would like to generate.
+ Choose the Inner Radius of generation.
+ Choose the Outer Radius of generation. (Max size of system)
+ Choose the Height of generation.
+ All other parameters should be acceptable.
-Use your custom ring configuration and generate the amount of mass / particles of each material type you would like. It is important to keep in mind how much mass you are generating, how much mass of EACH material type, how many particles and the mass of each particle, and the size of the system.
-Once you've generated all your materials to form the planets, you will notice nothing is happening.
-The way to spark the system is to:
+ Pause the simulation.
+ Select a particle / fragment.
+ Go to Actions and select "Convert Fragment to Full Body."
+ Repeat this for as many planetesimals as you would like. Spread them out unless you would like planets with moons.
+ Resume simulation.
NOTE: You may also use the ring generator, select "Bodies Instead of Particles", and choose the number of attracting bodies you would like. This may be more effective for determining the base material of formation and the starting mass of each planetesimal. I typically do not do this, as I like to look at the distribution of particles and determine clusters that would have a probability of clumping.
This method should be at least exciting, but that is barring the amount of simulation time it takes to get exciting results.
Play around with this and let me know what you think.