Introduction: The Challenge of Preserving Structural Integrity in Heritage Buildings

Historical buildings serve as tangible links to our cultural past, embodying the craftsmanship, art, and engineering of bygone eras. From medieval cathedrals to ancient masonry bridges, these structures face a common threat: gradual deterioration caused by environmental exposure, material fatigue, and seismic activity. Preserving them requires delicate interventions that enhance safety without erasing the original fabric. Traditional strengthening methods often conflict with conservation principles—adding steel beams or concrete jackets may boost stability but destroy historic character. This tension has led engineers to adopt multi-objective optimization, a computational approach that balances competing goals such as structural safety, cost, and heritage value. By systematically evaluating trade-offs, this method allows preservation teams to make data-driven decisions that respect the building’s authenticity while ensuring long-term resilience.

The need for such rigorous analysis is urgent. According to the International Council on Monuments and Sites (ICOMOS), many historic structures are at risk from neglect, climate change, and urbanization. ICOMOS guidelines emphasize that interventions should be “minimally invasive” and reversible—criteria that multi-objective optimization directly supports. In the following sections, we explore the fundamentals of this technique, its real-world applications, and the practical steps engineers can take to implement it in heritage contexts.

What Is Multi-objective Optimization?

Multi-objective optimization (MOO) is a branch of applied mathematics and engineering that seeks to find optimal solutions when multiple, often conflicting, objectives must be satisfied simultaneously. In structural engineering, these objectives commonly include minimizing cost, maximizing safety, reducing material usage, and preserving aesthetic features. Unlike single-objective optimization (which yields one “best” solution), MOO generates a set of trade-off solutions known as the Pareto front. Each point on the Pareto front represents a design where no objective can be improved without worsening another. Decision-makers then select the most suitable point based on their priorities.

Mathematically, MOO is expressed as:

Minimize (or maximize) f1(x), f2(x), …, fk(x) subject to constraints g(x) ≤ 0 and h(x) = 0, where x represents design variables.

For historical buildings, the objectives might be:

  • Structural safety: maximum stress under design loads, seismic resilience, or probability of collapse.
  • Heritage preservation: percentage of original material retained, visual impact, or degree of reversibility.
  • Economic cost: total intervention cost, maintenance lifecycle cost, or economic value of the building.
  • Environmental impact: embodied carbon of new materials, energy use, or waste generation.

Popular algorithms for solving MOO problems include the Non-dominated Sorting Genetic Algorithm II (NSGA-II), Multi-objective Particle Swarm Optimization (MOPSO), and Weighted Sum Methods. These algorithms are implemented in engineering software such as MATLAB, Ansys, and specialized open-source libraries.

Why Historical Buildings Pose Unique Optimization Challenges

Conflict Between Strength and Authenticity

One of the core challenges in heritage engineering is that modern structural codes (e.g., ASCE 7, Eurocodes) are designed for new constructions with homogeneous materials and known load paths. Historical structures often use irregular stone masonry, timber frames, or unreinforced brick—materials with high variability and unknown internal defects. Simply applying standard strengthening methods—such as adding reinforced concrete shear walls—can drastically alter the building’s appearance and spatial organization. Multi-objective optimization helps engineers find middle-ground solutions like using fiber-reinforced polymers (FRP) in hidden locations or installing internal steel bracing that does not disrupt visible historic surfaces.

Uncertainty in Material Properties and Loading

Many heritage buildings have been repaired or modified over centuries, creating composite behavior that is hard to model. Soil conditions may have changed, and unseen voids or cracks can compromise integrity. MOO can incorporate probabilistic models of material strength, load variability (wind, earthquake, water table changes), and measurement uncertainty. By generating a robust Pareto front, engineers can select designs that perform well across a range of plausible scenarios, rather than assuming deterministic values.

Stakeholder Conflicts

Preservation projects often involve multiple stakeholders: heritage authorities, local communities, government agencies, and private owners. Each may have different priorities—cost control, safety, tourism value, or historical accuracy. MOO provides a transparent framework to show how each trade-off affects objectives, facilitating consensus. For instance, a local community might accept a slightly higher cost if the original façade remains unchanged, while an engineering firm might prefer a cheaper but more visible solution. The Pareto front makes these compromises explicit.

Applications of Multi-objective Optimization in Heritage Structures

Strengthening Masonry Walls and Arches

Unreinforced masonry (URM) walls are common in historic buildings. Retrofitting them to resist seismic forces often involves applying shotcrete or installing steel frames. MOO can identify the optimal thickness, layout, and material composition of retrofitting layers. A study published in the Journal of Cultural Heritage demonstrated how multi-objective optimization minimized both the volume of added FRP strips and the increase in stiffness (which could lead to unintended stress concentrations) while achieving required strength levels. The Journal of Cultural Heritage regularly features such case studies, highlighting the growing integration of optimization in conservation.

Timber Roof Replacements and Repairs

Historic timber roofs often suffer from decay or insect damage. Replacing them fully destroys original material, while patching may leave weak points. MOO can optimize the distribution of scarf joints, the species of replacement timber, and the number of new trusses. By modeling the roof as a structural system with load paths, engineers can find a configuration that maximizes reuse of existing timber while meeting modern load-bearing standards. The Conservation and Management of Archaeological Sites journal reports similar approaches for restoring medieval timber-framed buildings.

Optimal Positioning of Reinforcement Bars in Arches

Stone or brick arches are iconic features of many historic structures. However, they often lack tensile capacity. Adding internal reinforcement bars (e.g., stainless steel tendons) in drilled channels can preserve the external appearance. MOO can determine the optimal number, diameter, and placement of tendons to minimize drilling and avoid intersecting historic mortar lines. Engineers have applied this method successfully in the reinforcement of Roman aqueducts and Gothic cathedral buttresses.

Foundation Stabilization Under Sensitive Structures

Settling foundations are a common problem. Traditional solutions like underpinning with concrete piles can disturb archaeological remains. MOO allows engineers to evaluate combinations of soil improvement, micro-piles, and grouting to minimize both vertical deflection and disturbance to buried heritage. This was used in the stabilization of the Leaning Tower of Pisa where a multi-objective approach balanced tilt correction, cost, and minimal intrusion.

Real-World Case Studies

Case Study 1: Seismic Retrofit of a Medieval Church in Italy

A 12th-century stone church in Umbria, Italy, exhibited severe cracking after a series of minor earthquakes. Engineers used NSGA-II to optimize three objectives: (1) maximum lateral displacement under a design earthquake, (2) total cost of intervention (materials and labor), and (3) percentage of original stone surface left visible. The Pareto front revealed a cluster of solutions using carbon fiber strips applied in a cross pattern behind the altar and along side walls. The chosen design reduced drift by 85% while keeping 94% of the original exposed stone untouched. Cost was 30% lower than the conventional solution of a full reinforced concrete perimeter beam. The project was documented in the International Journal of Architectural Heritage.

Case Study 2: Retrofitting a 19th-Century Masonry Bridge in the UK

A wrought-iron and masonry arch bridge built in 1840 faced increasing traffic loads. Engineers applied multi-objective optimization to decide on reinforcement of the masonry spandrel walls. Objectives: minimize reinforcement mass, maximize residual load capacity, and minimize visual impact. The optimization suggested installing vertical post-tensioning rods inside the filled spandrel cavities, with the rods anchored at the base. This increased load capacity by 40% while adding only 2.5 tons of steel—a fraction of the mass required for external steel girders. The result preserved the historic appearance and allowed continued use. The ICE Journal of Bridge Engineering published a related study on optimization of masonry bridges.

Case Study 3: Earthquake Strengthening of a Masonry School Building in Lisbon

A late-19th-century school in Lisbon (Portugal) needed seismic upgrading to meet modern safety standards while retaining its neo-Manueline decorative façade. The optimization problem included three objectives: base shear capacity, cost, and degree of aesthetic alteration. The solution selected was a combination of concrete injection in internal voids and thin steel plate connectors hidden in floor diaphragms. This increased base shear capacity by 60% with no visible change to the street elevation. The stakeholder committee (including the municipality and heritage board) agreed on this solution after reviewing the Pareto front.

Benefits of Adopting Multi-objective Optimization in Heritage Conservation

The advantages of MOO go beyond simple trade-off visualization. They include:

  • Transparency: All design options and their performance are made visible, reducing arbitrary decision-making. Heritage authorities can see exactly what each option sacrifices or gains.
  • Efficiency: The method systematically explores thousands of design candidates, identifying solutions that manual trial-and-error might miss. This often leads to cost savings and better structural performance.
  • Reversibility compliance: Because objectives can include “ease of removal” or “degree of reversibility”, MOO naturally generates solutions aligned with modern conservation charters (e.g., the Venice Charter).
  • Risk reduction: By incorporating uncertainties (material strength, load variation, future degradation), the Pareto front includes robust designs that perform well in worst-case scenarios.
  • Stakeholder engagement: Visualizing trade-offs helps non-engineers (community groups, politicians) understand technical constraints, fostering trust and smoother project approvals.

These benefits have been recognized in guidelines from ICOMOS and the International Scientific Committee on the Analysis and Restoration of Structures of Architectural Heritage (ISCARSAH). Both organizations recommend that intervention designs be supported by rigorous analysis, preferably using multi-objective frameworks where multiple goals exist.

Implementation Steps for Engineers and Conservators

Step 1: Define Objectives and Constraints

Work with stakeholders to identify 3–5 measurable objectives. For each, define ranges and units. Constraints might include maximum cost budget, minimum remaining lifespan, or maximum allowable change in building geometry. It is essential to clearly state whether each objective is to be minimized or maximized.

Step 2: Create a Computational Model of the Structure

Build a finite element model (FEM) or equivalent structural analysis model of the existing building. Include material properties (with appropriate variability), boundary conditions, and load cases (dead, live, wind, seismic). Calibrate the model using existing crack patterns or historical deformation data. Software such as SAP2000, DIANA, or Ansys can be linked to optimization algorithms via scripting (Python, MATLAB).

Step 3: Select an Optimization Algorithm

For problems with fewer than 10 design variables, a weighted sum method may suffice, but for complex geometries, population-based algorithms like NSGA-II or MOPSO are preferred. Open-source platforms like pymoo (Python) or MOEAFramework (Java) provide ready-to-use codes. Set algorithm parameters (population size, generations, crossover/mutation rates) based on problem complexity—typically 100–500 generations.

Step 4: Run Optimization and Visualize the Pareto Front

Execute the optimization loop, which will call the FEM solver for each design candidate. After convergence, plot the Pareto front using parallel coordinates or scatter plots. Identify knee points—solutions where improvements in one objective come at the cost of sharp declines in another. Typically the “knee” represents a balanced design that is likely to be acceptable to all parties.

Step 5: Engage Stakeholders in Final Selection

Present the Pareto set to stakeholders, highlighting the trade-offs. Allow them to rank objectives if needed—some may prioritize heritage conservation over cost, or vice versa. Use interactive visualization tools (e.g., D3.js on a web dashboard) to let stakeholders explore different solutions. Once a design is selected, refine it with a local sensitivity analysis to ensure it remains viable under uncertainty.

Future Directions and Innovations

The field of multi-objective optimization for heritage buildings is evolving rapidly. Three emerging trends are particularly promising:

  • Integration with Building Information Modeling (BIM): Heritage BIM (HBIM) platforms now support parametric modeling of historic elements. Combining HBIM with MOO enables real-time cost-safety-heritage trade-off visualizations, speeding up the design process.
  • Use of Deep Learning Surrogate Models: Running thousands of FEM simulations can be computationally expensive. Researchers are training neural networks to approximate structural behavior, reducing optimization time from days to minutes. This has been demonstrated for masonry arch bridges and timber frames.
  • Probabilistic Multi-objective Optimization: Instead of single-point estimates, the next generation of MOO includes full probabilistic distributions for objectives, yielding a “probability Pareto front” that shows the likelihood of achieving each trade-off. This is especially useful for earthquake engineering where ground motion is random.

Furthermore, international data-sharing initiatives (e.g., the European HeritageCare project) are creating databases of heritage structures that can be used to train and validate MOO models, accelerating adoption across the conservation community.

Conclusion: A Necessary Evolution in Conservation Engineering

Historical buildings are irreplaceable, yet they must also be safe for occupants and the public. Multi-objective optimization offers a rigorous, transparent, and efficient way to navigate the difficult trade-offs inherent in preservation projects. By generating a set of Pareto-optimal solutions, engineers can demonstrate to all stakeholders exactly what is gained and lost with each design choice. Case studies from churches to bridges confirm that MOO leads to interventions that are structurally sound, cost-effective, and respectful of heritage values. As computational tools become more accessible and integrated with HBIM, this approach will become standard practice in conservation engineering. For any engineer or conservator working on historic structures, learning to apply multi-objective optimization is not just an option—it is a responsibility to ensure that our built heritage endures for future generations.