docs on setting seed

Author: geraintpalmer

docs/Examples/mm1.rst

@@ -21,7 +21,6 @@ The following code repeats the experiment 100 times, only recording waits for th
 It then returns the average wait in the system::
     
     >>> import ciw
-    >>> from random import seed
     >>> def iteration(warmup):
     ...     N = ciw.create_network(params_dict)
     ...     Q = ciw.Simulation(N)
@@ -30,7 +29,7 @@ It then returns the average wait in the system::
     ...     waits = [row.waiting_time for row in records if row.arrival_date > warmup]
     ...     return sum(waits)/len(waits)
     
-    >>> seed(27)
+    >>> ciw.seed(27)
     >>> ws = []
     >>> for i in range(100):
     ...     ws.append(iteration(50))

docs/Features/deadlock.rst

@@ -40,8 +40,7 @@ Parameters::
 Running until deadlock::
 
     >>> import ciw
-    >>> from random import seed
-    >>> seed(99)
+    >>> ciw.seed(99)
     >>> N = ciw.create_network(params)
     >>> Q = ciw.Simulation(N, deadlock_detector='StateDigraph')
     >>> Q.simulate_until_deadlock()

docs/Features/index.rst

@@ -17,4 +17,5 @@ Contents:
    priority.rst
    deadlock.rst
    state_tracker.rst
-   exact.rst
+   exact.rst
+   seed.rst

docs/Features/seed.rst

@@ -0,0 +1,9 @@
+.. _setting-seeds:
+
+=============
+Setting Seeds
+=============
+
+The Ciw function :code:`ciw.seed` can be used to ensure reproducibility of results. As Ciw uses random number streams from both the :code:`random` and :code:`numpy` modules, this function simply sets the seeds for these.
+
+    >>> ciw.seed(5)