Multi-objective retrofit optimization model

Multi-objective retrofit optimization model#

Description

This analysis computes a series of linear programming models for single- and multi-objective optimization related to the effect of extreme weather on a community in terms of three objective functions. The three objectives used in this program are to minimize economic loss, minimize population dislocation, and maximize building functionality.

This analysis computes a series of linear programming models for single- and multi-objective optimization related to the effect of extreme weather on a community in terms of three objective functions. The three objectives used in this program are to minimize economic loss, minimize population dislocation, and maximize building functionality. The analysis uses the set of mitigation strategies, which is determined by the hazard type. For instance, seismic mitigation on existing buildings includes reinforcing buildings with cross bracing, reinforcing buildings using shear walls, install shear anchors, etc. For tsunami and flooding hazards, relocation is one of the mitigation strategies. Various parameters represent, for example, the starting and final retrofitting level of a building, the retrofitting cost for buildings retrofitted from one level to another in groups of structural types, or a coefficient of objective, which represents a community resilience goal to measure the performance of a system. The total number of objectives of the optimization model implemented in pyIncore is currently three; economic loss, population dislocation and building functionality constraints.

The computation proceeds by iteratively solving constrained linear models using epsilon steps. The CSV outputs of the computation are collections of optimal resource allocations.

Contributors

Science: Charles Nicholson and Yunjie Wen
Implementation: Dale Cochran, Tarun Adluri, Jorge Duarte, Diego Calderon, Santiago Núñez-Corrales, and NCSA IN-CORE Dev Team

Related publications

Input parameters

key name	type	name	description
`result_name`	`str`	Result name	Result CSV dataset name.
`model_solver`	`str`	Model solver	Choice of the model solver to use. Gurobi is the default solver.
`num_epsilon_steps` ^*	`int`	Epsilon values	Number of epsilon values to evaluate.
`max_budget` ^*	`str`	Maximum budget	Selection of maximum possible budget.
`budget_available`	`float`	Budget value	Custom budget value.
`inactive_submodels`	`List[int]`	Identifier of submodels	Identifier of submodels to inactivate during analysis.
`scale_data` ^*	`bool`	Scaling data	Choice for scaling data.
`scaling_factor`	`float`	Scaling factor	Custom scaling factor.

Input datasets

key name	type	name	description
`building_related_data` ^*	`incore:multiobjectiveBuildingRelatedData`	Building-related data	A csv file with building-related data.
`strategy_costs_data` ^*	`incore:multiobjectiveStrategyCost`	Strategy cost	A csv file with strategy cost data per building.

Output datasets

key name	type	parent key	name	description
`optimal_solution_dv_x` ^*	`incore:multiobjectiveOptimalSolutionX`		Results	A csv file with an optimal solution for decision variable x for Economic loss, Population dislocation and Building functionality values, results of objective function values.
`optimal_solution_dv_y` ^*	`incore:multiobjectiveOptimalSolutionY`		Results	A csv file with an optimal solution for decision variable y with initial and final retrofitted strategies for Economic loss, Population dislocation and Building functionality values, results of objective function values.

(* required)

Execution

code snippet:

    # Create Multi-objective retrofit optimization instance
    opt = `MultiObjectiveRetrofitOptimization`(client)

    # Load input dataset
    opt.load_remote_input_dataset("building_related_data", building_related_data)
    opt.load_remote_input_dataset("strategy_costs_data", strategy_costs_data)

    # Specify the result name
    result_name = "Multi_objective_retrofit_optimization"

    # Set analysis parameters
    opt.set_parameter("result_name", result_name)
    opt.set_parameter("model_solver", "Gurobi")
    opt.set_parameter("num_epsilon_steps", 10)
    opt.set_parameter("max_budget", max_budget)
    opt.set_parameter("budget_available", 1000000)
    opt.set_parameter("inactive_submodels", [inactive_submodels])
    opt.set_parameter("scale_data", True)
    opt.set_parameter("scaling_factor", 1.0)
    
    # Run Multi-objective retrofit optimization analysis
    opt.run_analysis()

full analysis: multi_retrofit_optimization.ipynb