Optimization Model Speeds Healthcare Provider Network Design
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current optimization software packages, such as Gurobi, require excessive computation time to find optimized healthcare provider networks, often taking over an hour to produce solutions, which is not feasible for real-time applications, especially when dealing with complex models involving 100,000 binary decision variables and constraints, and fails to meet the 10-minute time limit required for real-time network solution provision.
Innovation Solution
A computer-implemented method using a fractional mixed-integer linear optimization model enhanced with in-house computational innovations, specifically applying Benders decomposition and linear decomposition techniques to speed up the solution process without compromising solution quality, allowing for the generation of optimized provider networks within seconds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If leading commercial optimization software packages such as Gurobi are used to obtain optimized network solution, then solution quality is improved, but computation time increases to over an hour
Solution Approach 1:
The patent applies Benders decomposition to divide the complex optimization model into separate master problem and subproblems. The master problem handles strategic decisions about provider network selection, while subproblems evaluate operational feasibility and coverage requirements for each geographical designation. This segmentation allows each component to be solved more efficiently and iteratively, reducing overall computation time while maintaining solution quality.
Solution Approach 2:
The patent introduces an intermediary computational framework that acts as a mediator between the optimization model and the solution process. This framework includes custom preprocessing routines, iterative solution algorithms, and post-processing validation steps that efficiently bridge the gap between the complex model and real-time solution requirements, achieving both quality and speed.
2Measurement precision
If the optimization model involves 100,000 binary decision variables and model constraints to meet business requirements, then solution accuracy is improved, but model complexity increases
Solution Approach 1:
The patent segments the large-scale optimization model into manageable components through Benders decomposition. The master problem manages a reduced set of binary variables related to provider selection, while subproblems handle continuous variables related to coverage and capacity. This segmentation reduces the complexity of any single optimization problem while maintaining the integrity of the overall model and its 100,000 variables and constraints.
Solution Approach 2:
The patent transforms the complex high-dimensional optimization problem into a series of lower-dimensional problems solved iteratively. By moving between the master problem space (provider selection) and subproblem space (coverage verification), the solution approach navigates the complex solution space more efficiently, reducing the computational burden of directly solving the full 100,000-variable model.
3Productivity
If real-time network solution provision within 10-minute time limit is required, then responsiveness is improved, but solution quality may be compromised
Solution Approach 1:
The patent implements preliminary action through extensive preprocessing of the optimization model before the actual solution process. This includes pre-calculating coverage matrices, organizing provider data by geographical designation and specialty, and pre-defining constraint structures. These preliminary steps significantly reduce the computational burden during the actual optimization, enabling real-time solution delivery within the 10-minute limit while maintaining high solution quality.
Solution Approach 2:
The patent employs iterative solution algorithms that continuously refine the solution through alternating between master problem optimization and subproblem validation. This continuous action allows the system to progressively improve solution quality while monitoring computation time, ensuring that optimal or near-optimal solutions are achieved within the real-time 10-minute constraint rather than using static or approximate methods.
Data Source
AI summary
A computer-implemented method for selecting one or more providers having one or more specialties for a network available to members while satisfying one or more constraints, the method comprising: at a first computing device: receiving, from a second computing device separate and distinct from the first computer, a network provider request to select an optimized network of providers, the request including: a desired objective including at least one of provider cost minimization, provider average quality maximization and provider total volume maximization, one or more geographical designations, computing an optimized provider network of one or more selected providers including: applying a model that produces a non-optimized provider network of one or more providers having one or more specialties that satisfies the desired objective for the one or more geographical designations of the members without consideration of the availability in the geographical designation of one or more providers having one or more specialties, and applying a linear decomposition to the pre-optimized provider network to generate an optimized provider network; providing the optimized provider network to the second computing device.