Add solver list for more robust remote solving

[Fangop]

Dear @ds2010,

This patch provides a solver list for more robust remote solving.
Removing dependence on single solver of remote NEOS server makes the remote solving feature more robust.
However, I need helps for accomplish this enhancement.

Remote solving by MOSEK/CBC solvers:

• I tested a few instance solving with MOSEK and CBC with pyomo interface and NEOS server. All of these failed. I wonder do we have an example of successful execution which may help to enhance our package. (Maybe we did not pass the sufficient arguements?)

Complete solver list:

• Is there a function for etraction of LP solvers, MIP solvers, and NLP solvers in pyomo interface? It has a solver list like:
  ['bonmin', 'cbc', 'conopt', 'couenne', 'cplex', 'filmint', 'filter', 'ipopt', 'knitro', 'l-bfgs-b', 'lancelot', 'lgo', 'loqo', 'minlp', 'minos', 'minto', 'mosek', 'octeract', 'ooqp', 'path', 'raposa', 'snopt']
• Or, can you help me sort out different solvers? With more solvers in list, the less likely to fail optimization.

Thanks~
fix #54

[ds2010]

@Fangop Thanks for this pr. After NOES fixes this issue, I will review this pr then.

Let's wait a moment and see how we can enhance our package.

[Fangop]

Dear @ds2010 ,
This PR provides three functions for checking the local and remote solvers.
Though the implementation of checking solvers in the environment remains inpractical, these may enhance our package.

The check of the solver in the environment requires extra access to the system, which should not be the case with our package.

Here are some codes for testing the new feature:

from pystoned.utils.tools import check_local_solver, check_remote_solver, get_remote_solvers

print(check_local_solver("mosek"))
print(check_local_solver("gurobi"))
print(check_local_solver("ipopt"))

print(get_remote_solvers())

print(check_remote_solver("mosek"))
print(check_remote_solver("gurobi"))
print(check_remote_solver("ipopt"))