Machine Problem 4, due Apr 4

Part of the homework for CS:2820, Fall 2015
by Douglas W. Jones
THE UNIVERSITY OF IOWA Department of Computer Science

Write a Java program, in file NeuronNetwork.java that can read the description of a neuron network from a file called f (presuming that this file exists) when you type the following command to run your program:

[HawkID@serv15 project]$ java NeuronNetwork f

When your program is run, it should either:

Output error messages if the neuron-network description contains errors, or
Output a trace of the events that occur in a simulation of that neuron network, until there are no more events.
It should not output a copy of the simulation model (although you might want to keep the output code handy for debugging, do not include it in your final submission.

here is an example description (an extended version of the format used for MP3). This simple neuron network contains a clock that ticks periodically until a second neuron (responsible for counting the ticks) turns off the clock, causing activity in the network to terminate.

neuron CLOCK 10.0 11.0
synapse TICK CLOCK CLOCK 1.0 11.0

neuron COUNT 10.0 0.0
synapse - CLOCK COUNT 0.5 7.0
synapse - COUNT TICK  0.4 -0.6

Syntactically, there are two types of synapse, unnamed synapses and named ones. Synapse TICK above is a named synapse connecting neuron CLOCK back to itself. Unnamed synapses have a single dash in the name field. Neuron and synapse names be non-numeric, and each name may be defined only once.

Any synapse that is named may be the target of a secondary synapse. The synapse COUNT TICK above is a secondary synapse that connects neuron COUNT to (primary) synapse TICK. Logically (but not necessarily in your simulation), primary and secondary synapses are two distinct subclasses of synapse.

As before, the first number in the description of a neuron is its threshold, while the second number is its voltage at time zero.

As before, the first number in the description of a synapse is the delay, the second number is the strength. Delays may not be negative. Strengths are of either sign.

The simulation should implement the following model:

Whevever the voltage at a neuron exceeds the threshold of that neuron (including at the beginning of time), the neuron fires. When it fires, the voltage drops immediately to zero and an impulse is sent to each outgoing synapse.
Whenever an impulse is sent to a synapse, after the delay of that synapse, it will fire.
Whenever a primary synapse fires, the voltage at the destination neuron is incremented by the strength of that synapse (strengths may be negative for inhibitory synapses).
Neuron voltages decay with time. If the voltage at time t₁ was v₁, then if a synapse fires with strength s at time t₂ (and nothing else has happened to this neuron in the intervening time), the new voltage v₂ (prior to the neuron firing, if it does) is:
v₂ = v₁e^t₁–t₂ + s
Whenever a secondary synapse fires, the strength of the destination primary synapse is altered by adding the strength of the secondary synapse to it.

(This model leaves out a huge number of features of real neurons, but those features only add complexity to the details of events, they don't alter the structure of the simulation program. For example: In real life, the strength of a synapse is reduced each time it fires, and then the strength rebounds asymptotically towards its resting strength as the synapse recovers from firing. Another detail that could differ from neuron to neuron is the time constant of the delay.)

The output should show one line of text for each synapse that fires giving the time, the synapse name (actual name, if it has one, as well as source-destination), and the voltage of the destination neuron before and after the firing.

In addition, the output should show one line whenver a neuron fires, giving the time, the name of that neuron and saying that it fired.

An appropriate output from the simulation of the above model would begin something like the following and continue for several more time units before terminating:

0.0 neuron CLOCK
0.5 synapse - CLOCK COUNT 0.0 7.0
1.0 synapse TICK CLOCK CLOCK 0.0 11.0
1.0 neuron CLOCK

Grading criteria will be as usual: No credit will be given to programs that do not represent a substantial attempt at meeting the requirements of this assignment. Your program must:

Begin with a header comment giving the source file name.
Contain Javadoc comments claiming authorship and declaring the version to be MP4.
Be cleanly and readably indented.
Use appropriate variable names.
Respond appropriately to errors in the input.
Generate appropriate output if there were no errors in the input.
Not demonstrate any unhandled or mishandled exceptions.

Note that the requirements for header comments in the file header are absolute. Your name must appear in the form that it appears on your ID card. The TAs will not waste their time doing detective work to attempt to figure out who submitted what code.

Submission: Your solution should consist of a single file, with one public method, main. (Later, we will chop the file into smaller more managable pieces, as it grows.)

To submit your solution, make sure it is in a file called NeuronNetwork.java (this requires that the main method be in a class called NeuronNetwork. This must be in the current directory. Then, follow the following dialogue with the Linux shell (what you type is given in boldface):

[HawkID@serv16 ~/project]$ submit NeuronNetwork.java
Course (22C:016 would be c_016): CS2820
Possible submit directories for /group/submit/CS2820 are:

The dialogue continues as in MP1, except that the submit directory for this assignment will be named mp4 instead of mp1