In order to improve the operation of pneumatic transport systems, we developed two techniques that can be employed on existing systems to increase their efficiency. These techniques can be applied on-line to a system striving toward optimal performance. As an application of the technique, a 0.10-m diameter conveying line transporting 1 mm polyester particles over a 30-m distance was employed. The first approach to optimal performance uses either constant inlet pressure conditions or constant solids throughput. The iso-pressure drop method is performed by operating this system at two state parameters at a constant pressure while the constant solids throughput employs a similar procedure. The two state parameters are pressure drop and either the gas mass flow rate or solids rate. By individual changes in these parameters, one can explore small variations around an operating point and develop an approximation to the derivative of the parameter. Extrapolation of the information to the optimal condition can be carried out pursuing the point where the derivative goes to zero. Expanding the data set refines the operation and produces a more exact optimal condition. The second approach to optimal performance utilizes a neural network description of the conveying phase diagram. Once the description is defined, optimization procedures can be applied to hone in on the optimal operating point. This procedure involves training the neural network, which sometimes requires a large database for accuracy. The present application shows the speed with which the network can be trained.