Short answer: You'd need to simulate it with your choice of software.

Longer answer: Is this a Shipping container with RFID project? Why not install the antennas on the ceiling, where there's slightly less chance of blockage? Is the container full of something, because the contents will impact things.

Since it's a metal container, you can easily imagine the reflections from every surface. Due to lack of an exit, it's more like a microwave oven than a waveguide. You're using four antennas instead of a food turntable or old method stirrer. My gut instinct hints that a neat linear array isn't likely to be the optimum placement diversity, because in one dimension they're all somewhat in the same place.

.: You'll need software.

This is a large microwave oven problem. The RF energy bounces around if it's an empty container and you get hot and cold spots. If it's a full container the energy will bounce around alot if there is alot of metal inside, and the energy will get absorbed and weakened away from the antenna, and maybe not get to the far end, if the material is lossy, like water. 5x7x36 wavelength box is doable to analyze if you change the antenna to a waveguide port, I think. The accuracy will be bad most likely because analysis of not metal objects is based on the properties of the material.