11.3 Concurrent optimizer¶
The idea of the concurrent optimizer is to run multiple optimizations of the same problem simultaneously, and pick the one that provides the fastest or best answer. This approach is especially useful for problems which require a very long time and it is hard to say in advance which optimizer or algorithm will perform best.
The major applications of concurrent optimization we describe in this section are:
Using the interior-point and simplex optimizers simultaneously on a linear problem. Note that any solution present in the task will also be used for hot-starting the simplex algorithms. One possible scenario would therefore be running a hot-start simplex in parallel with interior point, taking advantage of both the stability of the interior-point method and the ability of the simplex method to use an initial solution.
Using multiple instances of the mixed-integer optimizer to solve many copies of one mixed-integer problem. This is not in contradiction with the run-to-run determinism of MOSEK if a different value of the MIO seed parameter
MSK_IPAR_MIO_SEEDis set in each instance. As a result each setting leads to a different optimizer run (each of them being deterministic in its own right).
The downloadable file contains usage examples of both kinds.
11.3.1 Common setup¶
We first define a method that runs a number of optimization tasks in parallel, using the standard multithreading setup available in the language. All tasks register for a callback function which will signal them to interrupt as soon as the first task completes successfully (with response code MSK_RES_OK).
here to download.¶# Defines a Mosek callback function whose only function
# is to indicate if the optimizer should be stopped.
stop = false
firstStop = 0
function callback(caller :: Callbackcode,
douinf :: Vector{Float64},
intinf :: Vector{Int32},
lintinf :: Vector{Int64})
if stop
1
else
0
end
end
When all remaining tasks respond to the stop signal, response codes and statuses are returned to the caller, together with the index of the task which won the race.
function runTask(num, task)
global stop
global firstStop
ttrm =
try
optimize(task)
catch e
if isa(e,MosekError)
return r.rcode,MSK_RES_ERR_UNKNOWN
else
rethrow()
end
end
# If this finished with success, inform other tasks to interrupt
# Note that data races around stop/firstStop are irrelevant
if ! stop
stop = true
firstStop = num
end
return MSK_RES_OK,ttrm
end
function optimizeconcurrent(tasks::Vector{Mosek.Task})
res = [ MSK_RES_ERR_UNKNOWN for t in tasks ]
trm = [ MSK_RES_ERR_UNKNOWN for t in tasks ]
# Set a callback function
for t in tasks
# Use remote server: putoptserverhost(task,"http://solve.mosek.com:30080")
putcallbackfunc(t, callback)
end
# Start parallel optimizations, one per task
Threads.@threads for i in 1:length(tasks)
(tres,ttrm) = runTask(i,tasks[i])
res[i] = tres
trm[i] = ttrm
end
# For debugging, print res and trm codes for all optimizers
for (i,(tres,ttrm)) in enumerate(zip(res,trm))
println("Optimizer $i res $tres trm $ttrm")
end
return firstStop, res, trm
end
11.3.2 Linear optimization¶
We use the multithreaded setup to run the interior-point and simplex optimizers simultaneously on a linear problem. The next methods simply clones the given task and sets a different optimizer for each. The result is the clone which finished first.
function optimizeconcurrent(task, optimizers)
# Choose various optimizers for cloned tasks
tasks = Mosek.Task[ let t = maketask(task)
# Use remote server: putoptserverhost(task,"http://solve.mosek.com:30080")
putintparam(t,MSK_IPAR_OPTIMIZER, opt)
t
end for opt in optimizers ]
# Solve tasks in parallel
firstOK, res, trm = optimizeconcurrent(tasks)
if firstOK > 0
return firstOK, tasks[firstOK], trm[firstOK], res[firstOK]
else
return 0, Nothing, Nothing, Nothing
end
end
It remains to call the method with a choice of optimizers, for example:
11.3.3 Mixed-integer optimization¶
We use the multithreaded setup to run many, differently seeded copies of the mixed-integer optimizer. This approach is most useful for hard problems where we don’t expect an optimal solution in reasonable time. The input task would typically contain a time limit. It is possible that all the cloned tasks reach the time limit, in which case it doesn’t really mater which one terminated first. Instead we examine all the task clones for the best objective value.
function optimizeconcurrentMIO(task, seeds)
# Choose various seeds for cloned tasks
tasks = Mosek.Task[ let t = maketask(task)
putintparam(MSK_IPAR_MIO_SEED, seed)
t
end for seed in seeds ]
# Solve tasks in parallel
(firstOK, res, trm) = optimizeconcurrent(tasks)
sense = getobjsense(task)
bestObj = if sense == MSK_OBJECTIVE_SENSE_MINIMIZE 1.0e+10 else -1.0e+10 end
bestPos = -1
if firstOK >= 0
# Pick the task that ended with res = ok
# and contains an integer solution with best objective value
for (i,t) in enumerate(tasks)
pobj = getprimalobj(t,MSK_SOL_ITG)
print("$i $pobj")
end
for (i,(tres,ttrm,t)) in enumerate(zip(res,trm,tasks))
solsta = getsolsta(t,MSK_SOL_ITG)
if tres == MSK_RES_OK &&
( solsta == MSK_SOL_STA_PRIM_FEAS ||
solsta == MSK_SOL_STA_INTEGER_OPTIMAL)
pobj = getprimalobj(t,MSK_SOL_ITG)
if ( ( sense == MSK_OBJECTIVE_SENSE_MINIMIZE &&
getprimalobj(t,MSK_SOL_ITG) < bestObj ) ||
( sense == MSK_OBJECTIVE_SENSE_MAXIMIZE &&
getprimalobj(t,MSK_SOL_ITG) > bestObj ) )
bestObj = pobj
bestPos = i
end
end
end
end
if bestPos > 0
return bestPos, tasks[bestPos], trm[bestPos], res[bestPos]
else
return 0, Nothing, Nothing, Nothing
end
end
It remains to call the method with a choice of seeds, for example:
