Hi Geoffrey,<div><br></div><div>Thanks for your reply.</div><div><br></div><div>> Does it make sense to wait longer than 7 mins after task A (presuming no other task forces occupies the user at that time)?<br>> Put differently: Can we say that the starting time of B = Math.max((endTime of task before B), (endTime of task A + 7 minutes))?<br>
> If we can say that, it's pointless to investigate the solutions where task B starts 8 minutes after task A and the user doing no task that last minute.<br></div><div><br></div><div>Yes, the starting time of B = Math.max((endTime of task before B), (endTime of task A + 7mins)) as long as it is smaller than (endTime of task A + 8mins).</div>
<div>Yes, it is pointless to investigate the solutions where task B starts 8 minutes after task A and the user doing no task that last minute. </div><div>The 8 minute is just a constraint that the task in between tasks A and B cannot take longer than 7:59s.</div>
<div><br></div><div>I am thinking that maybe instead of using time itself as the planning variable, we would use time just to determine the Hard and Soft scores. </div><div>So if Task B is scheduled after Task A + 8mins by the solver, then it inflicts on the hard score. Similarly if Task B is scheduled before Task A + 7 mins. </div>
<div>Does my reasoning make sense in any way?</div><div><br></div><div>thanks,</div><div>Josef</div><div><br></div><div><br><br><div class="gmail_quote">On 24 July 2012 20:46, Geoffrey De Smet <span dir="ltr"><<a href="mailto:ge0ffrey.spam@gmail.com" target="_blank">ge0ffrey.spam@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000">
<br>
<div>Op <a href="tel:23-07-12%2016" value="+35623071216" target="_blank">23-07-12 16</a>:26, Josef Bajada
schreef:<br>
</div><div class="im">
<blockquote type="cite">Hi Geoffrey,
<div><br>
</div>
<div>Well I want to leave 'space' between tasks in the situations
where there are hard constraints that require me to put this
space.<br>
</div>
</blockquote></div>
This makes the chaining technique harder to model, but I wouldn't
write it off yet.<div class="im"><br>
<blockquote type="cite">
<div><br>
</div>
<div>As a simple example:</div>
<div>
<br>
</div>
<div>Task A: Put pasta in boiling water (duration 40 seconds)</div>
<div>Task B: Take pasta out of boiling water (duration 50 seconds,
cannot start before 7 mins after Task A finishes, cannot start
after 8 mins after Task A finishes)</div>
</blockquote></div>
Does it make sense to wait longer than 7 mins after task A
(presuming no other task forces occupies the user at that time)?<br>
Put differently: Can we say that the starting time of B =
Math.max((endTime of task before B), (endTime of task A + 7
minutes))?<br>
If we can say that, it's pointless to investigate the solutions
where task B starts 8 minutes after task A and the user doing no
task that last minute.<br>
If we can say that, then chaining can calculate the the starting
time of a task on the fly differently.<div class="im"><br>
<blockquote type="cite">
<div>Task C: Chop vegetables (duration 2 minutes).</div>
<div><br>
</div>
<div>This will evidently leave some gaps. The ideal result from
the solver should be:</div>
<div><br>
</div>
<div>Task A: at time 0 (ends at 40s)</div>
<div>
Task C: at time 41s (ends at 2:41)</div>
<div>Task B: at time 7:40</div>
<div><br>
</div>
<div>There is a gap between C and B which is OK. </div>
<div><br>
</div>
<div>If another Task is added to the story:</div>
<div>
<div>Task D: Prepare sauce (duration 7 minutes)</div>
</div>
<div><br>
</div>
<div>I would want the following result:</div>
<div><br>
</div>
<div>Task A: at time 0 (ends at 40s)</div>
<div>Task D: at time 41s (ends 7:41s)</div>
<div>Task B: at time 8:42s (ends 9:32s)</div>
<div>Task C: at time 9:33s (ends 11:33s)</div>
<div><br>
</div>
<div>Task C can actually take place before Task A too. </div>
<div><br>
</div>
<div>I still need to read and understand the chaining
functionality properly. Do you think it would allow me to
achieve the above?</div>
<div><br>
</div>
</blockquote></div>
I don't know.<br>
But using continuous variables in a search problem such as this that
smells discrete with discrete constraints (A must start before B,
...), could blow up the search space unnecessarily.<br>
<br>
If you want to look into using continuous variables: the support for
it is limited currently.<br>
you can reuse the Drools Planner metaheuristic algorithms (including
termination, score, ...), but there's no decent generic move factory
support for continuous variables yet.<br>
So you 'll have to write a custom MoveFactory that creates a limited
subset of moves.<br>
Also, construction heuristics can't handle continuous variables yet,
so you 'll have to write a custom SolutionIntializer.<br>
There are examples with a custom MoveFactory and a custom
SolutionIntializer where you can copy paste from, but none with
continuous variables at the moment.<div><div class="h5"><br>
<blockquote type="cite">
<div>thanks,</div>
<div><br>
</div>
<div>Josef</div>
<div><br>
</div>
<div><br>
<br>
<div class="gmail_quote">On 22 July 2012 20:05, Geoffrey De Smet
<span dir="ltr"><<a href="mailto:ge0ffrey.spam@gmail.com" target="_blank">ge0ffrey.spam@gmail.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000"> Presuming that you
don't want to leave space between tasks, you can design
your model differently by using the "chained"
functionality:<br>
it will be far more efficient and the planning variable
won't be continuous.<br>
<br>
Let's presume you're scheduling Tasks to Persons.<br>
<br>
@PlanningEntity<br>
class Task implements TaskOrPerson {<br>
<br>
...<br>
<br>
@PlanningVariable(chained = true)<br>
@ValueRanges({<br>
@ValueRange(type =
ValueRangeType.FROM_SOLUTION_PROPERTY, solutionProperty =
"taskList"),<br>
@ValueRange(type =
ValueRangeType.FROM_SOLUTION_PROPERTY, solutionProperty =
"personList",<br>
excludeUninitializedPlanningEntity =
true)})<br>
public TaskOrPerson getPreviousTaskOrPerson() {<br>
return previousTaskOrPerson;<br>
}<br>
<br>
public int getDuration() {<br>
return duration;<br>
}<br>
<br>
public int getStartingTime() {<br>
int startingTime = 0;<br>
TaskOrPerson taskOrPerson =
getPreviousTaskOrPerson();<br>
while (taskOrPerson instanceof Task) { // Every
chain is guarantee to end up with an anchor (= Person)<br>
startingTime += ((Task)
taskOrPerson).getDuration();<br>
taskOrPerson = ((Task)
taskOrPerson).getPreviousTaskOrPerson()<br>
}<br>
return startingTime;<br>
}<br>
<br>
}<br>
<br>
class Person implements TaskOrPerson {<br>
<br>
}<br>
<br>
For a good example, take a look at the VehicleRouting
example.<br>
For more info about chaining, in the manual see section
4.3.4.2.6. Chained<br>
<a href="http://docs.jboss.org/drools/release/5.4.0.Final/drools-planner-docs/html_single/index.html" target="_blank">http://docs.jboss.org/drools/release/5.4.0.Final/drools-planner-docs/html_single/index.html</a><br>
<br>
<div>Op <a href="tel:22-07-12%2018" value="+35622071218" target="_blank">22-07-12 18</a>:00, Josef Bajada
schreef:<br>
</div>
<blockquote type="cite">
<div>
<div>Hi,
<div><br>
</div>
<div>I am new to Drools and Drools Planner, so
apologies if I am asking anything obvious.</div>
<div><br>
</div>
<div>My objective is to implement a simple (for now)
planner which schedules tasks according to 2 main
criteria:</div>
<div>- Their duration (in seconds)</div>
<div>- Their dependencies on other tasks (e.g. Hard
Constraint that Task B has to start between 180
and 200 seconds after Task A finishes).</div>
<div><br>
</div>
<div>Since there are gaps between dependent tasks as
part of the hard constraints other tasks can be
fitted in between dependent tasks.</div>
<div>So the Solver needs to find the optimal start
time for each task that satisfies the hard
constraints, and in the shortest total timeline
possible to complete all tasks (soft constraint).</div>
<div><br>
</div>
<div>The main problem I am finding is that this
start time, which is essentially the planning
variable is a continuous variable. </div>
<div>Chapter 4 of the Drools documentation mentions
very briefly (Section 4.3.4.1) that planning
variables can be continuous, but there does not
seem to be any more details about how to achieve
this.</div>
<div><br>
</div>
<div>Even if the planning variable was discrete (say
bins of 5 second intervals), there is no upper
bound as such.</div>
<div><br>
</div>
<div>How is it best to handle such planning
variables in Drools Planner?</div>
<div><br>
</div>
<div>thanks,<br>
josef</div>
<div><br>
</div>
<div><br>
</div>
<br>
<fieldset></fieldset>
<br>
</div>
</div>
<pre>_______________________________________________
rules-users mailing list
<a href="mailto:rules-users@lists.jboss.org" target="_blank">rules-users@lists.jboss.org</a>
<a href="https://lists.jboss.org/mailman/listinfo/rules-users" target="_blank">https://lists.jboss.org/mailman/listinfo/rules-users</a><span><font color="#888888">
</font></span></pre>
<span><font color="#888888"> </font></span></blockquote>
<span><font color="#888888"> <br>
<pre cols="72">--
With kind regards,
Geoffrey De Smet</pre>
</font></span></div>
<br>
_______________________________________________<br>
rules-users mailing list<br>
<a href="mailto:rules-users@lists.jboss.org" target="_blank">rules-users@lists.jboss.org</a><br>
<a href="https://lists.jboss.org/mailman/listinfo/rules-users" target="_blank">https://lists.jboss.org/mailman/listinfo/rules-users</a><br>
<br>
</blockquote>
</div>
<br>
</div>
<br>
<fieldset></fieldset>
<br>
<pre>_______________________________________________
rules-users mailing list
<a href="mailto:rules-users@lists.jboss.org" target="_blank">rules-users@lists.jboss.org</a>
<a href="https://lists.jboss.org/mailman/listinfo/rules-users" target="_blank">https://lists.jboss.org/mailman/listinfo/rules-users</a>
</pre>
</blockquote>
<br>
<pre cols="72">--
With kind regards,
Geoffrey De Smet</pre>
</div></div></div>
<br>_______________________________________________<br>
rules-users mailing list<br>
<a href="mailto:rules-users@lists.jboss.org">rules-users@lists.jboss.org</a><br>
<a href="https://lists.jboss.org/mailman/listinfo/rules-users" target="_blank">https://lists.jboss.org/mailman/listinfo/rules-users</a><br>
<br></blockquote></div><br></div>