CHOOSE
Energy Portfolio Optimization
Determine the optimal long-term mix of thermal, renewable, and storage assets to minimize cost while meeting reliability and emissions targets over 10-30 year planning horizons.
Understanding the Problem
Electric utilities face an unprecedented generation planning challenge. Over the next 10-30 years, they must completely reimagine their asset portfolios to meet three simultaneous imperatives: accommodate explosive demand growth driven by electrification and data centers, maintain reliability as baseload thermal plants retire, and achieve deep decarbonization targets. The U.S. Department of Energy projects approximately 104 GW of coal and natural gas retirements by 2030, offset by 209 GW of new capacity, yet only 10% of new additions will be firm baseload power. The business stakes are enormous. Poor planning decisions can result in stranded assets, reliability failures, regulatory penalties, or massive cost overruns passed to ratepayers. Traditional fragmented planning methods across generation, transmission, and distribution silos lead to suboptimal decisions that can cost billions. Conversely, utilities that master integrated portfolio optimization can identify least-cost pathways that maintain reliability while achieving emissions targets, quantify the value of investment flexibility, and build stakeholder confidence through transparent, defensible infrastructure plans.
THE CHALLENGE
What Makes it Hard
Utilities must determine which power plants to build, maintain, or retire over multi-decade horizons while balancing cost minimization, reliability standards, and emissions reduction targets. This requires evaluating thousands of potential investment pathways across uncertain futures.
WHO FACES IT
Variable renewable integration creates massive computational challenges in modeling wind and solar intermittency, grid stability, and the loss of stabilizing rotating mass as thermal plants retire, causing frequency and voltage volatility.
Transmission-generation timeline mismatch creates planning complexity: solar and batteries deploy in months, but transmission lines require 10+ years of planning and permitting, risking stranded assets and delayed interconnection.
Multi-objective optimization under deep uncertainty requires balancing cost, reliability, and emissions across decades of unknown fuel prices, technology costs, demand growth, and evolving regulations with computationally intractable full temporal resolution.
BUSINESS IMPACT
Optimize billions in energy infrastructure: achieve 20% O&M cost reduction while navigating the shift to renewables (93% of global capacity additions).
O&M Cost Reduction
20%
Potential Savings[1]
North American utilities need to reduce O&M costs by more than $15 billion (approximately 20%) over five years to meet earnings targets.
Renewable Capacity
93%
Global Additions[2]
Renewables accounted for 92.5% of total global capacity expansion in 2024, representing 585 GW of new renewable capacity.
Emissions Reduction
95%+
Lower vs Coal[3]
Renewable energy sources emit approximately 30-50g CO2eq/kWh compared to 1,000g for coal, a 95%+ reduction in emissions intensity.
How We Solve It
We solve generation portfolio optimization using stochastic mixed-integer linear programming with advanced decomposition methods. Benders decomposition separates investment decisions from operational dispatch, while representative day sampling makes multi-decade horizons tractable. Scenario analysis across hundreds of futures identifies robust strategies.
Hybrid Compute
What We Bring
Multi-decade capacity expansion modeling with hourly dispatch resolution
Stochastic optimization across hundreds of fuel price, demand, and policy scenarios
Hybrid solver orchestration combining MILP solvers with quantum-inspired methods
Seamless integration with industry-standard tools like ReEDS, GenX, and switch-model
Real-time sensitivity analysis for regulatory proceedings and stakeholder presentations
FUTURE POSSIBILITIES
The
Quantum Horizon
Generation portfolio optimization presents a theoretically interesting combinatorial structure with tens of thousands of binary build/retire decisions. While current quantum hardware cannot handle production-scale problems, the discrete asset selection subproblems map naturally to QUBO formulations being explored by quantum researchers.
Exploratory Work
A proof-of-concept would focus on a single-year capacity expansion subproblem with 50-100 candidate assets, comparing quantum annealing and QAOA results against classical MILP. This demonstrates the formulation approach while acknowledging that production-scale problems require classical decomposition methods.
Current Research Directions
Quantum annealing for simplified single-period capacity expansion via QUBO formulation on D-Wave systems
Hybrid quantum-classical decomposition separating discrete investment decisions from continuous dispatch
Quantum-inspired evolutionary algorithms (NSGA-II, MOEA/D) for multi-objective portfolio optimization
Digital annealing approaches using Fujitsu hardware for large-scale scenario exploration
Interested in quantum research?
Explore proof-of-concept implementations with our team.
Ready to solve this problem?
Talk to our experts about how Strangeworks can help with energy portfolio optimization.
SOURCES