Two-Stage Feedstock Procurement Optimization Under Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for optimizing feedstock procurement in process industries, such as refineries, fail to effectively hedge against uncertainties in market and operation conditions, often relying on deterministic techniques that do not account for real-world decision-making steps and are computationally impractical for complex models.
Innovation Solution
A two-stage approach using chance-constrained optimization for long-term contracts and breakeven analysis for short-term spot market decisions, simulating multiple outcomes based on uncertain input parameters to determine robust feedstock selection and procurement volumes, maintaining computationally tractable models without complex decomposition strategies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deterministic techniques are used for feedstock procurement optimization, then the model is computationally simple and easy to solve, but it fails to hedge against uncertainties in market and operation conditions
Solution Approach 1:
The patent segments the feedstock procurement optimization into two distinct planning steps: long-term contract decisions and short-term spot market decisions. Each step handles specific types of uncertainties appropriately, with the first stage focusing on market conditions and the second stage incorporating operation conditions. This segmentation allows the model to address uncertainties without requiring a single overly complex formulation.
Solution Approach 2:
The patent changes the mathematical formulation from deterministic to probabilistic by introducing chance-constrained optimization in the first stage and breakeven analysis in the second stage. This parameter change enables the model to hedge against uncertainties by incorporating probability distributions of uncertain parameters while maintaining computational tractability through efficient solution algorithms.
2Reliability
If stochastic programming is used to handle uncertainties, then the model can account for multiple scenarios, but it becomes computationally impractical for complex refinery models
Solution Approach 1:
The patent divides the stochastic optimization problem into two separate planning stages, each with its own uncertainty considerations and solution methods. The first stage uses chance-constrained optimization for long-term contracts with fewer scenarios, while the second stage uses breakeven analysis for spot market decisions. This segmentation dramatically reduces computational complexity compared to a single-stage stochastic program while still capturing the essential uncertainties.
Solution Approach 2:
The patent extracts and separates the handling of different types of uncertainties into distinct methodological components. Rather than incorporating all uncertainties into a single complex stochastic program, the patent extracts market uncertainties for the first stage and operation uncertainties for the second stage, applying appropriate solution methods to each. This extraction maintains reliability while improving computational efficiency.
3Reliability
If long-term contract decisions are made months in advance with uncertain spot market knowledge, then feed availability and cost assurance are improved, but the ability to profit from spot market upsides is reduced
Solution Approach 1:
The patent segments procurement into long-term contracts and spot market purchases, with each serving distinct strategic purposes. The long-term contract decisions provide the hedging foundation for feed availability and cost assurance, while the subsequent spot market decisions enable flexibility to capture market upsides. This temporal and functional segmentation resolves the contradiction by allowing both objectives to be pursued in sequence.
Solution Approach 2:
The patent incorporates feedback from the first stage results into the second stage decision-making. The long-term contract allocations determined in stage one inform the spot market procurement opportunities evaluated in stage two. This feedback mechanism allows the system to maintain the stability provided by long-term contracts while adapting to spot market conditions, thus achieving both cost assurance and profit opportunities.
4Adaptability or versatility
If spot market decisions are made in the short-term, then the ability to capture market upsides is improved, but consideration of operation condition uncertainties must be increased
Solution Approach 1:
The patent segments the decision-making process so that operation condition uncertainties are primarily addressed in the second stage through breakeven analysis, rather than complicating the first stage. This segmentation allows spot market decisions to focus on capturing opportunities while systematically incorporating operation uncertainties through the breakeven framework, maintaining clarity and reducing overall decision complexity.
Solution Approach 2:
The patent performs preliminary analysis of operation conditions and equipment availability before making spot market decisions. By preparing the breakeven analysis framework and assessing operational constraints in advance of spot market trading, the system can quickly evaluate spot opportunities without ad hoc complexity. This preliminary action reduces the decision complexity during actual spot market execution.
Data Source
AI summary
A computer system and method optimize feedstock selection planning for an industrial process by evaluating first and second stages at separate intervals throughout the planning process. Evaluating the first stage determines a set of robust feedstocks to procure on long-term contracts. The computer system and method solve, in parallel, multiple simulation cases of a non-linear model generated with different expectation values for uncertain input parameters related to selecting feedstocks to procure on long-term contracts. Probabilistic analyses on the solutions from the simulation cases, including the application of chance-constraints, determine the set of robust feedstocks to procure on long-term contracts. Evaluating the second stage determines a set of robust feedstocks to procure in the spot market, using the information from the first stage. Specifically, the computer system and method solve each of multiple new simulation cases of the non-linear model, generated with different expectation values for uncertain input parameters related to selecting feedstocks to procure in the spot market. Each simulation case is solved to determine breakeven prices for one or more available spot feedstocks, and probabilistic analyses are performed on the breakeven prices for these spot feedstocks to determine a set of robust feedstocks to procure in the spot market.