ficomo 0.1.2

ficomo: a pyomo-like framework for the FICO xpress solver

---

ficomo: a pyomo-like framework for the FICO xpress solver

ficomo is a library that provides a set of tools for using the FICO xpress solver in the same way as pyomo. ficomo supports the following features:

Features

✨ Declarative syntax: ficomo uses a declarative syntax for defining variables, expressions and constraints.

🚀 Native support for FICO xpress: ficomo provides a native interface to the xpress solver.

🔍 Diagnostics: ficomo provides tools for analyzing optimization problems, such as finding infeasible constraints, finding relaxations that will make the problem feasible, "force" solving an infeasible problem, etc.

⛓️ Remote execution: ficomo provides an easy way to run optimization problems on a remote executor node in both sync and async modes.

💾 Save and load: ficomo provides a way to save and load optimization problems and solutions without depending on xpress implementation details.

Usage Example

Here's a basic example of using ficomo to solve a simple optimization problem. The model determines the optimal number of chairs and tables to produce given manufacturing time and storage space constraints, while maximizing profit.

from ficomo.environ import Problem, Remote, SearchFor, Sense, var

USE_REMOTE = False # change to True to use remote execution
REMOTE_URL = None # None runs a local executor and connects to it

# Define decision variables
x1 = var(name="chairs", lb=0)  # number of chairs to produce
x2 = var(name="tables", lb=0)  # number of tables to produce

# Define objective: maximize profit
# Profit = 20 per chair + 30 per table
objective = 20 * x1 + 30 * x2

# Add constraints
# Resource constraints: time available for manufacturing
# Each chair takes 4 hours, each table takes 6 hours
# Total time available is 48 hours
time_constraint = 4 * x1 + 6 * x2 <= 48

# Storage space constraint:
# Each chair takes 0.2 sq m, each table takes 0.4 sq m
# Available space is 5 sq m
space_constraint = 0.2 * x1 + 0.4 * x2 <= 5.0

# Create the problem
variables = [x1, x2]
constraints = [time_constraint, space_constraint]
problem = Problem(variables, constraints, objective, sense=Sense.maximize)

if USE_REMOTE:
    # Solve the problem on a remote executor node
    with Remote(REMOTE_URL) as remote:
        solution = problem.to(remote).solve(search_for=SearchFor.optimal)

else:
    # Solve the problem localy
    solution = problem.solve(search_for=SearchFor.optimal)

# Print solution status
print(f"Solve status: {solution.solve_status!r}")
print(f"Solution status: {solution.solution_status!r}")

# Print solution
print(f"Optimal production plan: {solution.value({x1: x1, x2: x2})}")
print(f"Total profit: ${solution.value(objective)}")

# Output:
# Solve status: <SolveStatus.COMPLETED: 3>
# Solution status: <SolutionStatus.OPTIMAL: 1>
# Optimal production plan: {chairs: 12.0, tables: 0.0}
# Total profit: $240.0

# Save the problem to production-problem.fxp
problem.dump("production-problem.fxp")

# Save the solution to solution.fxs
# NOTE: the solution contains the problem
solution.dump("production-plan.fxs")

# Load the problem from production-problem.fxp
# problem = Problem.load("production-problem.fxp")

# Load the solution from solution.fxs
# solution = Solution.load("production-plan.fxs")