This PR introduces a new module for representing (nonlinear) input/output systems and support for simulation, linearization, and composition. The following features are supported:

• Input/output systems can be created using the LinearIOSystem and NonlinearIOSystem classes, and I/O systems can be simulated in continuous and discrete time using the input_output_response() function, which is equivalent to the forced_response() function for LTI system.

• Input/output systems can be composed using the existing series, parallel, and feedback functions, as well as using the * and + operators. In addition, the InterconnectedSystem class can be used to interconnect a collection of I/O subsystems, with functionality similar to the connect function in MATLAB.

• Input/output systems and input/output signals can be named and Interconnections between I/O systems can be specified using system and signal names.

• Input/output systems can be linearized about an equilibrium point using the linearize() function, which returns a LinearIOSystem object. The LinearIOSystem is derived from the StateSpace class, allowing all operations on linear systems to be applied to the linearized system. The find_eqpt function can be used to find an equilibrium point for an input/output system.

• Documentation and examples are included to illustrate the use of the iosys module.

For the most part, the changes in this PR are restricted to the new iosys.py module and the corresponding test and documentation files. Some small modifications were required in other modules to support the features of this new module.

This PR is an initial release that implements fairly minimal functionality. A list of additional features that should be added are included at the top of the iosys.py file.