Methods for improved production and distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing optimization methods for air separation plant networks struggle to efficiently solve the combined production and distribution optimization problem due to high variability in electricity costs, complex production and distribution dynamics, and the inability to incorporate new data in a timely manner, leading to sub-optimal solutions.
Innovation Solution
A modified genetic algorithm approach that discretizes plant operation modes and customer sourcing decisions, allowing for the incorporation of intermediate data, thereby reducing the problem size and enabling quick decision-making by focusing on discrete variables and binary operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional continuous optimization methods are used to solve the combined production and distribution problem, then solution accuracy is improved, but solution time becomes excessively long and computational complexity increases
Solution Approach 1:
The patent segments the continuous production and distribution optimization problem into discrete time periods and discrete decision variables. By dividing the problem into manageable discrete segments, the optimization can be solved efficiently using genetic algorithms while maintaining practical applicability for production scheduling and distribution planning.
Solution Approach 2:
The patent transforms the continuous optimization parameters into discrete parameters suitable for genetic algorithm processing. Production rates, distribution quantities, and timing decisions are converted from continuous variables to discrete alternatives, enabling computational efficiency while preserving the essential decision-making structure of the problem.
2Adaptability or versatility
If all possible plant sources are considered in the distribution optimization, then customer sourcing options are improved, but problem size becomes exorbitant and difficult to solve
Solution Approach 1:
The patent segments the set of all possible plant sources into meaningful groups or categories based on product type, geographic region, or operational characteristics. This segmentation reduces the effective problem size by allowing the optimization to process structured subsets of sources while still considering comprehensive sourcing options across all segments.
Solution Approach 2:
The patent implements a two-stage optimization approach where a preliminary filtering stage identifies a reduced set of relevant plant sources based on basic criteria, and then the full optimization is applied only to this reduced set. This partial action approach maintains solution quality while dramatically reducing computational complexity.
3Loss of energy
If electricity price variability is fully incorporated into production decisions, then cost optimization is improved, but decision-making complexity increases
Solution Approach 1:
The patent performs preliminary analysis of electricity price patterns and forecasts to identify optimal production windows before making detailed production decisions. By pre-processing the price variability data and identifying key decision points in advance, the system reduces the complexity of real-time decision-making while capturing the cost optimization opportunities from price variability.
Solution Approach 2:
The patent structures production decisions around periodic electricity price cycles, aligning production schedules with predictable price patterns. By organizing decisions around regular periodic cycles rather than continuous price fluctuations, the system simplifies decision-making while still exploiting price variability for cost optimization.
Data Source
AI summary
A computer-implemented system and method for producing and distributing at least one product from at least one plant to at least one customer where discretized plant production data, filtered customer sourcing data, forecasted customer demand data, and forecasted plant electricity pricing data are input into a modified genetic algorithm and an electronic processor solves the modified genetic algorithm and outputs the solution to an interface. The system and method is flexible and can incorporate data as it becomes available to yield intermediate solutions for quick decision making.